All Products
Search
Document Center

Elastic Compute Service:Available instance families

Last Updated:Jun 22, 2026

This topic provides an overview of ECS instance families, including their features, instance types, and use cases, to help you select the right instance type for your needs.

Note

Catalog

Enterprise-level x86 instance families

General-purpose (g-series)

Intel processor

AMD processor

Not recommended (If unavailable, use a recommended instance family)

Compute-optimized (c-series)

Intel processor

AMD processor

Not recommended (If unavailable, use a recommended instance family)

Memory-optimized (r-series)

Intel processor

AMD processor

Not recommended (If unavailable, use a recommended instance family)

Universal (u-series)

Universal instance family u2a

Universal instance family u2i

Universal instance family u1

Big data (d-series)

Recommended

Not recommended (If unavailable, use a recommended instance family)

Local SSD (i-series)

Intel® Xeon® Granite Rapids processors

Intel® Xeon® Scalable processors (Ice Lake)

Intel®Xeon® Platinum 8269CY (Cascade Lake)

Intel®Xeon® Platinum 8163 (Skylake)

High clock speed (hf-series)

Recommended

Not recommended (If unavailable, use a recommended instance family)

Intel® Xeon® 6 processors with P-cores

4th Generation Intel® Xeon® Scalable processors (Sapphire Rapids)

Intel®Xeon® Cooper Lake processors

Intel®Xeon® Platinum 8269CY (Cascade Lake)

Enhanced

Storage-enhanced

Network-enhanced

Security-enhanced

Memory-enhanced

Entry-level x86 instance families

Recommended

Not recommended

Enterprise-level Arm compute instance families

Yitian 710 processor

Ampere® Altra® processor

ECS Bare Metal instance families

Recommended

Not recommended

General-purpose (g)

Compute-optimized (c)

Memory-optimized (r)

High clock speed (hf)

GPU-accelerated compute-optimized (gn)

SCC instance families

Heterogeneous computing instance families

Recommended

Not recommended

Enterprise X86 compute instance family

g9ae general-purpose instance family

  • Introduction: g9ae instances are built on the new Cloud Infrastructure Processing Unit (CIPU) architecture. They are equipped with the latest AMD EPYC Turin processors that use physical cores. These instances provide stable computing power, a powerful I/O engine, and chip-level security hardening.

  • Scenarios: Big data analytics (such as Spark, Flink, and ES), search, recommendation, and advertising (ps-worker), core transaction systems, audio and video transcoding, AI training and inference, and general-purpose enterprise applications (Java).

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: AMD EPYC Turin processor with a turbo frequency of up to 3.7 GHz. These processors use physical cores to provide stable computing performance.

    • For information about operating system compatibility, see AMD instance types and operating system compatibility.

  • Storage:

  • Network:

    • You can adjust the baseline network bandwidth.

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security:

The g9ae instance familyincludes the instance types and metrics listed in the following table.

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline disk IOPS

Baseline/Burst disk bandwidth (Gbit/s)

ecs.g9ae.large

2

8

2.5/Up to 25

Up to 1.5 million

Up to 0.5 million

2

3

6

6

Up to 200,000

2.5/Up to 20

ecs.g9ae.xlarge

4

16

4/Up to 25

Up to 1.6 million

Up to 0.5 million

4

4

15

15

Up to 200,000

3/Up to 20

ecs.g9ae.2xlarge

8

32

6/Up to 25

Up to 2.5 million

Up to 0.5 million

8

4

15

15

Up to 200,000

4/Up to 20

ecs.g9ae.4xlarge

16

64

10/Up to 25

Up to 3.2 million

Up to 0.5 million

16

8

30

30

Up to 200,000

5.5/Up to 20

ecs.g9ae.8xlarge

32

128

16/Up to 25

Up to 5 million

Up to 1 million

32

8

30

30

Up to 200,000

8/Up to 20

ecs.g9ae.12xlarge

48

192

25 / None

7.5 million

1.5 million

48

8

30

30

150,000

13/N/A

ecs.g9ae.16xlarge

64

256

32 or None

10 million

2 million

64

8

30

30

200,000

16 / None

ecs.g9ae.24xlarge

96

384

50 / None

15 million

3 million

64

15

30

30

300,000

25/N/A

ecs.g9ae.32xlarge

128

512

64/N/A

20 million

4 million

64

15

30

30

400,000

32 or None

ecs.g9ae.48xlarge

192

768

100 or Unlimited

30 million

6 million

64

15

50

50

600,000

50 / None

g9a general-purpose instance family

  • Introduction: g9a instances are built on the new Cloud Infrastructure Processing Unit (CIPU) architecture. They are equipped with the latest AMD EPYC Turin processors. These instances provide stable computing power, a powerful I/O engine, and chip-level security hardening.

  • Scenarios: Large and medium-sized database systems, game servers, financial quantization, blockchain, web and application servers, and other general-purpose enterprise applications.

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: AMD EPYC Turin processor with a turbo frequency of up to 4.1 GHz. These processors provide stable computing performance.

    • For information about operating system compatibility, see AMD instance types and operating system compatibility.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: These instances support the Trusted Platform Module (vTPM) feature.

The g9a instance family includes the following instance types and metrics.

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Cloud Disk IOPS Basics

Baseline/Burst disk bandwidth (Gbit/s)

ecs.g9a.large

2

8

2.5/Up to 15

Up to 1.2 million

Up to 0.5 million

2

3

6

6

Up to 110,000

2/Up to 15

ecs.g9a.xlarge

4

16

4/Up to 15

Up to 1.4 million

Up to 0.5 million

4

4

15

15

Up to 110,000

3/Up to 15

ecs.g9a.2xlarge

8

32

6/Up to 15

Up to 2 million

Up to 0.5 million

8

4

15

15

Up to 110,000

4/Up to 15

ecs.g9a.4xlarge

16

64

12/Up to 25

Up to 3 million

Up to 0.5 million

16

8

30

30

Up to 110,000

5/Up to 15

ecs.g9a.8xlarge

32

128

16/Up to 32

Up to 4 million

Up to 0.8 million

32

8

30

30

Up to 110,000

8/Up to 15

ecs.g9a.16xlarge

64

256

32/None

7.5 million

1.5 million

64

8

30

30

120,000

16 or None

ecs.g9a.24xlarge

96

384

48/N/A

10 million

2.2 million

64

15

30

30

200,000

24 / None

ecs.g9a.32xlarge

128

512

64 / None

15 million

3 million

64

15

30

30

250,000

32/None

ecs.g9a.48xlarge

192

768

96 or None

20 million

4.5 million

64

15

50

50

400,000

48 / None

g9i general-purpose instance family

  • Introduction: g9i instances are built on the new Cloud Infrastructure Processing Unit (CIPU) architecture. They are equipped with Intel® Xeon® 6 processors that feature P-cores (performance cores). These instances provide stable computing power, a powerful I/O engine, and chip-level security hardening.

  • Scenarios: High packet forwarding rate scenarios, game servers, small and medium-sized database systems, caches, search clusters, search and promotion applications, web and application servers, data analytics and computing, and trusted and secure computing scenarios.

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: Intel® Xeon® Granite Rapids with a clock speed of 3.2 GHz, an all-core turbo frequency of 3.6 GHz, and a single-core maximum turbo frequency of 3.9 GHz.

      Note

      The system may display different frequencies for this instance. The single-core maximum turbo frequency of 3.9 GHz is a burst performance. The burst capability depends on the overall CPU load of the physical machine and is not covered by the Service-Level Agreement (SLA).

    • For information about operating system compatibility, see Intel instance types and operating system compatibility.

  • Storage:

    • These are I/O optimized instances.

    • These instances support the NVMe protocol. For more information, see Overview of the NVMe protocol.

    • Supported disk types: ESSDs, ESSD AutoPL disks, and Regional ESSDs. For more information, see Overview of Elastic Block Storage.

    • For small instance types, disk IOPS and disk bandwidth are burstable. The storage I/O performance of an instance corresponds to its instance type. A larger instance type provides higher storage I/O performance. For more information, see Storage I/O performance.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

g9i includes the instance types and metrics listed in the following table:

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline/Burst disk IOPS

Baseline/Burst disk bandwidth (Gbit/s)

ecs.g9i.large

2

8

2.5/Up to 15

1 million

Up to 0.5 million

2

3

6

6

25,000/Up to 200,000

2/Up to 10

ecs.g9i.xlarge

4

16

4/Up to 15

1.2 million

Up to 0.5 million

4

4

15

15

50,000/Up to 200,000

2.5/Up to 10

ecs.g9i.2xlarge

8

32

6/Up to 15

1.6 million

Up to 0.5 million

8

4

15

15

60,000/Up to 200,000

4/Up to 10

ecs.g9i.3xlarge

12

48

10/Up to 15

2.4 million

Up to 0.5 million

12

8

15

15

80,000/Up to 200,000

5/Up to 10

ecs.g9i.4xlarge

16

64

12/Up to 25

3 million

0.5 million

16

8

30

30

100,000/Up to 200,000

6/Up to 10

ecs.g9i.6xlarge

24

96

15/Up to 25

4.5 million

0.6 million

24

8

30

30

120,000/Up to 200,000

7.5/Up to 10

ecs.g9i.8xlarge

32

128

20/Up to 32

6 million

0.8 million

32

8

30

30

200,000/Up to 300,000

10/Up to 12

ecs.g9i.12xlarge

48

192

25/Up to 32

9 million

1.6 million

48

8

30

30

240,000/Up to 320,000

12/Up to 15

ecs.g9i.16xlarge

64

256

28/Up to 36

12 million

2 million

64

8

30

30

300,000/Up to 400,000

16/Up to 24

ecs.g9i.24xlarge

96

384

32/Up to 48

18 million

3 million

64

15

30

30

350,000/Up to 600,000

20/Up to 28

ecs.g9i.32xlarge

128

512

36/Up to 50

20 million

4 million

64

15

30

30

400,000/Up to 650,000

24/Up to 28

ecs.g9i.48xlarge

192

768

64/N/A

24 million

6 million

64

15

50

50

500,000/Up to 800,000

32 or None

g8a general-purpose instance family

  • Family introduction: This instance family is built on the new Alibaba Cloud Cloud Infrastructure Processing Unit (CIPU) architecture and provides stable computing power, a more powerful I/O engine, and chip-level security hardening.

  • Scenarios: General-purpose enterprise applications (Java), memory-optimized or relational database applications, big data applications (such as Kafka and Elasticsearch), web applications, AI training and inference, and audio and video transcoding applications.

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: AMD EPYC Genoa 9T24 processor with a base frequency of 2.7 GHz and a turbo frequency of up to 3.7 GHz. These processors provide stable computing performance.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

    • For information about operating system compatibility, see AMD instance types and operating system compatibility.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • These instances provide an ultra-high packet forwarding rate.

    • For small instance types, network bandwidth is burstable.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

The g8a instance family includes the instance types and metric data listed in the following table.

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline/Burst disk IOPS

Baseline/Burst disk bandwidth (Gbit/s)

ecs.g8a.large

2

8

1.5/12.5

0.9 million

Up to 0.25 million

2

3

6

6

20,000/110,000

1.5/10

ecs.g8a.xlarge

4

16

2.5/12.5

1 million

Up to 0.25 million

4

4

15

15

30,000/110,000

2/10

ecs.g8a.2xlarge

8

32

4/12.5

1.6 million

Up to 0.25 million

8

4

15

15

45,000/110,000

2.5/10

ecs.g8a.4xlarge

16

64

7/12.5

2 million

0.3 million

16

8

30

30

60,000/110,000

3.5/10

ecs.g8a.8xlarge

32

128

10/25

3 million

0.6 million

32

8

30

30

80,000/110,000

5/10

ecs.g8a.12xlarge

48

192

16/25

4.5 million

0.75 million

48

8

30

30

120,000 or None

8/10

ecs.g8a.16xlarge

64

256

20/25

6 million

1 million

64

8

30

30

160,000/Unlimited

10/None

ecs.g8a.24xlarge

96

384

32/None

9 million

1.5 million

64

15

30

30

240,000 or Unlimited

16/None

ecs.g8a.32xlarge

128

512

40 or None

12 million

2 million

64

15

30

30

320,000/Unlimited

20/None

ecs.g8a.48xlarge

192

768

64 or None

18 million

3 million

64

15

50

50

500,000/None

32/None

Note
  • The packet forwarding rate (PPS) varies significantly based on the business scenario. Therefore, perform stress testing on your services to understand the performance of instances and select an appropriate instance type.

  • For ecs.g8a.large and ecs.g8a.xlarge instances, you must enable jumbo frames to achieve the burst network bandwidth of 12.5 Gbit/s. For more information, see Jumbo frames.

g8i general-purpose instance family

  • Family introduction: Built on the new Alibaba Cloud Cloud Infrastructure Processing Unit (CIPU) architecture, this family provides stable computing power, a more powerful I/O engine, and chip-level security hardening.

  • Scenarios: High packet forwarding rate scenarios, game servers, small and medium-sized database systems, caches, search clusters, search and promotion applications, web and application servers, data analytics and computing, and trusted and secure computing scenarios.

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: Intel® Xeon® Emerald Rapids or Intel® Xeon® Sapphire Rapids with a clock speed of at least 2.7 GHz and an all-core turbo frequency of 3.2 GHz. These processors provide stable computing performance.

      Note

      When you purchase an instance of this family, the system randomly assigns one of the preceding processors. You cannot manually select a processor.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

    • For information about operating system compatibility, see Intel instance types and operating system compatibility.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security:

The g8i instance family includes the instance types and metrics that are listed in the following table.

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline/Burst disk IOPS

Baseline/Burst disk bandwidth (Gbit/s)

ecs.g8i.large

2

8

2.5/Up to 15

1 million

Up to 0.3 million

2

3

6

6

25,000/Up to 200,000

2/Up to 10

ecs.g8i.xlarge

4

16

4/Up to 15

1.2 million

Up to 0.3 million

4

4

15

15

50,000/Up to 200,000

2.5/Up to 10

ecs.g8i.2xlarge

8

32

6/Up to 15

1.6 million

Up to 0.3 million

8

4

15

15

60,000/Up to 200,000

4/Up to 10

ecs.g8i.3xlarge

12

48

10/Up to 15

2.4 million

Up to 0.3 million

12

8

15

15

80,000/Up to 200,000

5/Up to 10

ecs.g8i.4xlarge

16

64

12/Up to 25

3 million

0.35 million

16

8

30

30

100,000/Up to 200,000

6/Up to 10

ecs.g8i.6xlarge

24

96

15/Up to 25

4.5 million

0.5 million

24

8

30

30

120,000/Up to 200,000

7.5/Up to 10

ecs.g8i.8xlarge

32

128

20/Up to 25

6 million

0.8 million

32

8

30

30

200,000/Unlimited

10/None

ecs.g8i.12xlarge

48

192

25/N/A

9 million

1 million

48

8

30

30

300,000/No limit

12/N/A

ecs.g8i.16xlarge

64

256

32/None

12 million

1.6 million

64

8

30

30

360,000 / None

20 / None

ecs.g8i.24xlarge

96

384

50/none

18 million

2 million

64

15

30

30

500,000/Unlimited

24 / None

ecs.g8i.48xlarge

192

1024

100/None

30 million

4 million

64

15

50

50

1,000,000 / Unlimited

48/None

g8ine network-enhanced general-purpose instance family

  • Family introduction: Powered by Alibaba Cloud's new Cloud Infrastructure Processing Unit (CIPU) architecture, this family delivers stable computing power and a more powerful I/O engine.

  • Scenarios: Network-intensive scenarios that require excellent forwarding performance and a high number of connections. These instances are especially suitable for network access layer gateways and traffic and data forwarding or pre-processing middleware. They can be used as part of a cloud solution in scenarios such as large websites, e-commerce, and AI.

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: Intel® Xeon® Emerald Rapids or Intel® Xeon® Sapphire Rapids with a clock speed of at least 2.7 GHz and an all-core turbo frequency of 3.2 GHz. These processors provide stable computing performance.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

  • Storage:

    • These are I/O optimized instances.

    • These instances support the NVMe protocol. For more information, see Overview of the NVMe protocol.

    • Supported disk types: ESSDs, ESSD AutoPL disks, and regional ESSDs. For more information, see Overview of Elastic Block Storage.

    • For small instance types, disk IOPS and disk bandwidth are burstable.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

The g8ine instance family includes the instance types and metrics described in the following table.

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

EBS Multi-queue

Baseline/Burst disk IOPS

Baseline/Burst disk bandwidth (Gbit/s)

ecs.g8ine.large

2

8

4/Up to 24

0.6 million

2

3

10

10

1

20,000/Up to 80,000

2/Up to 8

ecs.g8ine.xlarge

4

16

7/Up to 28

1.2 million

4

4

15

15

1

40,000/Up to 80,000

2.5/Up to 8

ecs.g8ine.2xlarge

8

32

12/Up to 35

2 million

8

6

15

15

2

50,000/Up to 80,000

4/Up to 8

ecs.g8ine.4xlarge

16

64

23/Up to 44

3.5 million

16

8

30

30

2

80,000/Up to 100,000

6/Up to 10

ecs.g8ine.8xlarge

32

128

44 / None

7 million

32

8

30

30

4

100,000 / Unlimited

10 or None

g8ae performance-enhanced general-purpose instance family

  • Introduction: g8ae instances are built on the new Cloud Infrastructure Processing Unit (CIPU) architecture. They provide stable computing power, a powerful I/O engine, and chip-level security hardening.

  • Scenarios: Artificial intelligence (AI) scenarios (such as deep learning, training, and AI inference), high-performance computing (HPC) and other high-performance scientific computing scenarios, large and medium-sized database systems, caches, search clusters, large online game servers, and other general-purpose enterprise applications that have high performance requirements.

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: AMD EPYC™ Genoa processor with a clock speed of 3.4 GHz and a single-core turbo frequency of up to 3.75 GHz. These processors provide stable computing performance.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

    • For information about operating system compatibility, see AMD instance types and operating system compatibility.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • These instances provide an ultra-high packet forwarding rate.

    • For small instance types, network bandwidth is burstable.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

The g8ae instance family includes the instance types and metrics listed in the following table.

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

vTPM support

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline/Burst disk IOPS

Baseline/Burst disk bandwidth (Gbit/s)

ecs.g8ae.large

2

8

3/Up to 15

1 million

Yes

Up to 0.3 million

2

3

6

6

30,000/Up to 200,000

2/Up to 10

ecs.g8ae.xlarge

4

16

4/Up to 15

1.2 million

Yes

Up to 0.3 million

4

4

15

15

50,000/Up to 200,000

2.5/Up to 10

ecs.g8ae.2xlarge

8

32

6/Up to 15

1.6 million

Yes

Up to 0.3 million

8

4

15

15

60,000/Up to 200,000

3/Up to 10

ecs.g8ae.4xlarge

16

64

12/Up to 25

3 million

Yes

0.5 million

16

8

30

30

100,000/Up to 200,000

6/Up to 10

ecs.g8ae.8xlarge

32

128

20/Up to 25

6 million

Yes

1 million

32

8

30

30

200,000 or Unlimited

10 or None

ecs.g8ae.16xlarge

64

256

32/None

9 million

Yes

1.5 million

64

8

30

30

250,000 or Unlimited

16/None

ecs.g8ae.32xlarge

128

512

64 / None

18 million

Yes

3 million

64

15

30

30

500,000/None

32 / None

Note

For ecs.g8ae.large and ecs.g8ae.xlarge instances, you must enable jumbo frames to achieve the burst bandwidth of 15 Gbit/s. For more information, see Jumbo frames.

g7a general-purpose instance family

  • Introduction: Built on the third-generation SHENLONG architecture, g7a instances provide stable and predictable ultra-high performance. They also use chip-level acceleration to provide significant improvements in storage and network performance, and computing stability.

  • Scenarios: Video encoding and decoding, high packet forwarding rate scenarios, web and application servers, small and medium-sized database systems, caches, search clusters, game servers, test and development (such as DevOps), and other general-purpose enterprise applications.

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: 2.55 GHz AMD EPYC MILAN processor with a single-core turbo frequency of up to 3.5 GHz. These processors provide stable computing performance.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

    • For information about operating system compatibility, see AMD instance types and operating system compatibility.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances provide an ultra-high packet forwarding rate.

    • For small instance types, network bandwidth is burstable.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

The g7a instance familyincludes the following instance types and metrics.

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline/Burst disk IOPS

Baseline/Burst disk bandwidth (Gbit/s)

ecs.g7a.large

2

8

1/Up to 10

0.9 million

Up to 0.25 million

2

3

6

6

12,500/Up to 110,000

1/Up to 6

ecs.g7a.xlarge

4

16

1.5/Up to 10

1 million

Up to 0.25 million

4

4

15

15

20,000/Up to 110,000

1.5/Up to 6

ecs.g7a.2xlarge

8

32

2.5/Up to 10

1.6 million

Up to 0.25 million

8

4

15

15

30,000/Up to 110,000

2/Up to 6

ecs.g7a.4xlarge

16

64

7/Up to 12.5

2 million

0.3 million

8

8

30

30

60,000/Up to 110,000

3.7/Up to 10.5

ecs.g7a.8xlarge

32

128

8/Up to 10

3 million

0.6 million

16

7

30

30

75,000/Up to 110,000

4.1/Up to 11

ecs.g7a.16xlarge

64

256

16 or None

6 million

1 million

32

8

30

30

150,000/Unlimited

8.2/N/A

ecs.g7a-nps1.16xlarge

64

256

16 or None

6 million

1 million

32

8

30

30

150,000/Unlimited

8.2/None

ecs.g7a.32xlarge

128

512

32 or None

12 million

2 million

32

15

30

30

300,000/Unlimited

16.4/N/A

Note

The kernels of the Ubuntu 16 and Debian 9 operating systems do not support AMD EPYC MILAN processors. Therefore, do not use Ubuntu 16 or Debian 9 images to create instances of this family. Otherwise, the instances fail to start.

g7, general-purpose instance family

  • Introduction: Built on the third-generation SHENLONG architecture, g7 instances provide stable and predictable ultra-high performance. They also use chip-level acceleration to provide significant improvements in storage and network performance, and computing stability.

  • Scenarios: High packet forwarding rate scenarios (such as live comments and telecommunication service forwarding), game servers, small and medium-sized database systems, caches, search clusters, enterprise applications of various types and sizes, web and application servers, data analytics and computing, trusted and secure computing scenarios, and blockchain scenarios.

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: Third-generation Intel® Xeon® Scalable (Ice Lake) processor with a base frequency of 2.7 GHz and an all-core turbo frequency of 3.5 GHz. These processors provide stable computing performance.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • These instances provide an ultra-high packet forwarding rate.

    • For small instance types, network bandwidth is burstable.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security:

The g7 instance family includes the instance types and performance metrics that are listed in the following table.

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

vTPM support

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Maximum number of attached data disks

Baseline/Burst disk IOPS

Baseline/Burst disk bandwidth (Gbit/s)

ecs.g7.large

2

8

2/Up to 12.5

1.1 million

Yes

Up to 0.5 million

2

3

6

6

8

20,000/Up to 160,000

1.5/Up to 10

ecs.g7.xlarge

4

16

3/Up to 12.5

1.1 million

Yes

Up to 0.5 million

4

4

15

15

8

40,000/Up to 160,000

2/Up to 10

ecs.g7.2xlarge

8

32

5/Up to 15

1.6 million

Yes

Up to 0.5 million

8

4

15

15

16

50,000/Up to 160,000

3/Up to 10

ecs.g7.3xlarge

12

48

8/Up to 15

2.4 million

Yes

Up to 0.5 million

8

8

15

15

16

70,000/Up to 160,000

4/Up to 10

ecs.g7.4xlarge

16

64

10/Up to 25

3 million

Yes

0.5 million

8

8

30

30

16

80,000/Up to 160,000

5/Up to 10

ecs.g7.6xlarge

24

96

12/Up to 25

4.5 million

Yes

0.55 million

12

8

30

30

16

110,000/Up to 160,000

6/10

ecs.g7.8xlarge

32

128

16/Up to 32

6 million

Yes

0.6 million

16

8

30

30

24

160,000 / No limit

10 or None

ecs.g7.16xlarge

64

256

32/None

12 million

Yes

1.2 million

32

8

30

30

32

360,000 or None

16/None

ecs.g7.32xlarge

128

512

64/None

24 million

Yes

2.4 million

32

15

30

30

32

600,000 or unlimited

32/None

g6 general-purpose instance family

  • Introduction: Built on the SHENLONG architecture, g6 instances offload many virtualization features to dedicated hardware. This reduces virtualization overhead and provides stable and predictable ultra-high performance.

  • Scenarios:

    • High packet forwarding rate scenarios, such as live comments and telecommunication service forwarding.

    • Enterprise applications of various types and sizes.

    • Web and application servers.

    • Game servers.

    • Small and medium-sized database systems, caches, and search clusters.

    • Data analytics and computing.

    • Compute clusters and memory-dependent data processing.

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: 2.5 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processor with a turbo frequency of 3.2 GHz. These processors provide stable computing performance.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

    Note

    This family does not support FreeBSD 13.2 and earlier versions due to compatibility issues. Use FreeBSD 13.3 or a later version.

  • Storage:

    • These are I/O optimized instances.

    • Supported disk types: ESSDs, ESSD AutoPL disks, Regional ESSDs, standard SSD, and ultra disk. For more information, see Block Storage Overview.

      Note

      The maximum performance of disks varies based on the instance families. A single instance of this instance family can deliver up to 200,000 IOPS.

    • The storage I/O performance of an instance corresponds to its instance type. A larger instance type provides higher storage I/O performance. For more information, see Storage I/O performance.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances provide an ultra-high packet forwarding rate.

      Note

      Network performance varies based on the instance families. For higher concurrent connection and network packet forwarding capabilities, we recommend that you use the g7ne instance family.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

The g6 instance family includes the instance types and metrics listed in the following table.

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.g6.large

2

8

1/Up to 3

0.3 million

Up to 0.25 million

2

2

6

1

10,000

1

ecs.g6.xlarge

4

16

1.5/Up to 5

0.5 million

Up to 0.25 million

4

3

10

1

20,000

1.5

ecs.g6.2xlarge

8

32

2.5/Up to 8

0.8 million

Up to 0.25 million

8

4

10

1

25,000

2

ecs.g6.3xlarge

12

48

4/Up to 10

0.9 million

Up to 0.25 million

8

6

10

1

30,000

2.5

ecs.g6.4xlarge

16

64

5/Up to 10

1 million

0.3 million

8

8

20

1

40,000

3

ecs.g6.6xlarge

24

96

7.5/Up to 10

1.5 million

0.45 million

12

8

20

1

50,000

4

ecs.g6.8xlarge

32

128

10 or None

2 million

0.6 million

16

8

20

1

60,000

5

ecs.g6.13xlarge

52

192

12.5 or None

3 million

0.9 million

32

7

20

1

100,000

8

ecs.g6.26xlarge

104

384

25 / None

6 million

1.8 million

32

15

20

1

200,000

16

g6a general-purpose instance family

  • Family introduction: Built on the SHENLONG architecture, this family offloads numerous virtualization features to dedicated hardware, which reduces virtualization overhead and provides stable and predictable ultra-high performance.

  • Scenarios:

    • Video encoding and decoding.

    • Scenarios with high packet forwarding rates.

    • Web and application servers.

    • Small and medium-sized database systems, caches, and search clusters.

    • Game servers.

    • Test and development, such as DevOps.

    • Other general-purpose enterprise applications.

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: 2.6 GHz AMD EPYC ROME processor with a turbo frequency of 3.3 GHz. These processors provide stable computing performance.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

    • For information about operating system compatibility, see AMD instance types and operating system compatibility.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances provide an ultra-high packet forwarding rate.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

The g6a instance family includes the instance types and metric data shown in the following table.

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.g6a.large

2

8

1/10

0.9 million

Up to 0.25 million

2

2

6

1

12,500

1

ecs.g6a.xlarge

4

16

1.5/10

1 million

Up to 0.25 million

4

3

15

1

20,000

1.5

ecs.g6a.2xlarge

8

32

2.5/10

1.6 million

Up to 0.25 million

8

4

15

1

30,000

2

ecs.g6a.4xlarge

16

64

5/10

2 million

0.3 million

8

8

30

1

60,000

3.1

ecs.g6a.8xlarge

32

128

8/10

3 million

0.6 million

16

7

30

1

75,000

4.1

ecs.g6a.16xlarge

64

256

16 or None

6 million

1 million

32

8

30

1

150,000

8.2

ecs.g6a.32xlarge

128

512

32 or None

12 million

2 million

32

15

30

1

300,000

16.4

g6e performance-enhanced general-purpose instance family

  • Introduction: Built on the third-generation SHENLONG architecture, g6e instances offload many virtualization features to dedicated hardware. This reduces virtualization overhead and provides stable and predictable ultra-high performance. They also use chip-level acceleration to provide significant improvements in storage and network performance, and computing stability.

  • Scenarios:

    • High packet forwarding rate scenarios, such as live comments and telecommunication service forwarding.

    • Enterprise applications of various types and sizes.

    • Web and application servers.

    • Game servers.

    • Small and medium-sized database systems, caches, and search clusters.

    • Data analytics and computing.

    • Compute clusters and memory-dependent data processing.

  • Compute:

    • Processor-to-memory ratio of approximately 1:4.

    • Processor: 2.5 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processor with a turbo frequency of 3.2 GHz. These processors provide stable computing performance.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

    Note

    This instance family does not support FreeBSD 13.2 or earlier versions due to compatibility issues. You can use FreeBSD 13.3 or later.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances provide an ultra-high packet forwarding rate.

      Note

      Network performance varies based on the instance families. For higher concurrent connection and network packet forwarding capabilities, we recommend that you use the g7ne instance family.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

The g6e instance family includes the instance types and metrics described in the following table.

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.g6e.large

2

8

1.2/Up to 10

0.9 million

Up to 0.25 million

2

3

6

1

20,000

1

ecs.g6e.xlarge

4

16

2/Up to 10

1 million

Up to 0.25 million

4

4

15

1

40,000

1.5

ecs.g6e.2xlarge

8

32

3/Up to 10

1.6 million

Up to 0.25 million

8

4

15

1

50,000

2

ecs.g6e.4xlarge

16

64

6/Up to 10

3 million

0.3 million

8

8

30

1

80,000

3

ecs.g6e.8xlarge

32

128

10 / None

6 million

0.6 million

16

8

30

1

150,000

5

ecs.g6e.13xlarge

52

192

16 / None

9 million

1 million

32

7

30

1

240,000

8

ecs.g6e.26xlarge

104

384

32 or None

24 million

1.8 million

32

15

30

1

480,000

16

Note
  • Network capabilities represent the maximum performance in single-item tests. For example, when testing network bandwidth, stress tests are not simultaneously performed on the packet forwarding rate or other metrics.

  • To use ecs.g6e.26xlarge, you must submit a ticket to request access.

g5 general-purpose instance family

  • Scenarios:

    • High packet forwarding rate scenarios, such as live comments and telecommunication service forwarding.

    • Enterprise applications of various types and sizes.

    • Small and medium-sized database systems, caches, and search clusters.

    • Data analytics and computing.

    • Compute clusters and memory-dependent data processing.

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: 2.5 GHz Intel® Xeon® Platinum 8163 (Skylake) or 8269CY (Cascade Lake) processor that provides stable computing performance.

      Note

      Instances of this family may be deployed on different server platforms. If your business requires instances to be deployed on the same server platform, we recommend that you use g6, g6e, or g7 instances.

  • Storage:

    • These are I/O optimized instances.

    • Supported disk types: ESSDs, ESSD AutoPL disks, standard SSD, and ultra disk. For more information, see Block Storage Overview.

      Note

      The maximum performance of disks varies based on the instance families. A single instance of this instance family can deliver up to 200,000 IOPS.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances provide an ultra-high packet forwarding rate.

      Note

      Network performance varies based on the instance families. For higher concurrent connection and network packet forwarding capabilities, we recommend that you use the g7ne instance family.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

The g5 instance family includes the following instance types and metrics.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Packets per second (PPS)

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

ecs.g5.large

2

8

1

0.3 million

2

2

6

1

ecs.g5.xlarge

4

16

1.5

0.5 million

2

3

10

1

ecs.g5.2xlarge

8

32

2.5

0.8 million

4

4

10

1

ecs.g5.3xlarge

12

48

4

0.9 million

4

6

10

1

ecs.g5.4xlarge

16

64

5

1 million

4

8

20

1

ecs.g5.6xlarge

24

96

7.5

1.5 million

6

8

20

1

ecs.g5.8xlarge

32

128

10

2 million

8

8

20

1

ecs.g5.16xlarge

64

256

20

4 million

16

8

20

1

sn2ne network-enhanced general-purpose instance family

  • Scenarios:

    • High packet forwarding rate scenarios, such as live comments and telecommunication service forwarding.

    • Enterprise applications of various types and sizes.

    • Small and medium-sized database systems, caches, and search clusters.

    • Data analytics and computing.

    • Compute clusters and memory-dependent data processing.

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: 2.5 GHz Intel® Xeon® E5-2682 v4 (Broadwell), Platinum 8163 (Skylake), or 8269CY (Cascade Lake) processor that provides stable computing performance.

      Note

      Instances of this family may be deployed on different server platforms. If your business requires instances to be deployed on the same server platform, we recommend that you use g6, g6e, or g7 instances.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances provide an ultra-high packet forwarding rate.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

The sn2ne instance family includes the instance types and metrics listed in the following table.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

ecs.sn2ne.large

2

8

1

0.3 million

2

2

6

1

ecs.sn2ne.xlarge

4

16

1.5

0.5 million

2

3

10

1

ecs.sn2ne.2xlarge

8

32

2

1 million

4

4

10

1

ecs.sn2ne.3xlarge

12

48

2.5

1.3 million

4

6

10

1

ecs.sn2ne.4xlarge

16

64

3

1.6 million

4

8

20

1

ecs.sn2ne.6xlarge

24

96

4.5

2 million

6

8

20

1

ecs.sn2ne.8xlarge

32

128

6

2.5 million

8

8

20

1

ecs.sn2ne.14xlarge

56

224

10

4.5 million

14

8

20

1

Compute-optimized instance family c9ae

  • Introduction: Built on the new-generation Alibaba Cloud CIPU and powered by the latest AMD EPYC Turin processors, these instances use physical cores to deliver stable computing performance, a powerful I/O engine, and chip-level security hardening.

  • Use cases: Ideal for workloads such as big data analytics (e.g., Spark, Flink, and Elasticsearch), search and recommendation engines, advertising systems (ps-worker), core transaction systems, audio and video transcoding, AI training and inference, and general-purpose enterprise applications (e.g., Java).

  • Compute:

    • Processor-to-memory ratio: 1:2.

    • Processor: AMD EPYC Turin processors with a turbo frequency of up to 3.7 GHz.

    • For information about OS compatibility, see AMD instance type and OS compatibility.

  • Storage:

  • Network:

  • Security:

The following table lists the available c9ae instance types and their specifications.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

NIC queues

ENIs

Private IPv4 addresses

IPv6 addresses

Disk IOPS

Disk bandwidth (Gbit/s)

ecs.c9ae.large

2

4

2.5 / 25

Up to 1,500,000

Up to 500,000

2

3

6

6

Up to 200,000

2.5 / 20

ecs.c9ae.xlarge

4

8

4 / 25

Up to 1,600,000

Up to 500,000

4

4

15

15

Up to 200,000

3 / 20

ecs.c9ae.2xlarge

8

16

6 / 25

Up to 2,500,000

Up to 500,000

8

4

15

15

Up to 200,000

4 / 20

ecs.c9ae.4xlarge

16

32

10 / 25

Up to 3,200,000

Up to 500,000

16

8

30

30

Up to 200,000

5.5 / 20

ecs.c9ae.8xlarge

32

64

16 / 25

Up to 5,000,000

Up to 1,000,000

32

8

30

30

Up to 200,000

8 / 20

ecs.c9ae.12xlarge

48

96

25 / None

7,500,000

1,500,000

48

8

30

30

150,000

13 / None

ecs.c9ae.16xlarge

64

128

32 / None

10,000,000

2,000,000

64

8

30

30

200,000

16 / None

ecs.c9ae.24xlarge

96

192

50 / None

15,000,000

3,000,000

64

15

30

30

300,000

25 / None

ecs.c9ae.32xlarge

128

256

64 / None

20,000,000

4,000,000

64

15

30

30

400,000

32 / None

ecs.c9ae.48xlarge

192

384

100 / None

30,000,000

6,000,000

64

15

50

50

600,000

50 / None

Compute-optimized instance family c9a

  • Introduction: Built on the new-generation Alibaba Cloud CIPU architecture and powered by the latest AMD EPYC Turin processors to provide stable computing performance, a powerful I/O engine, and chip-level security hardening.

  • Use cases: Ideal for workloads such as large and medium-sized database systems, game servers, financial quantitative analysis, blockchain, web and application servers, and other general-purpose enterprise applications.

  • Compute:

    • The vCPU-to-memory ratio is 1:2.

    • Processor: AMD EPYC Turin processor with a turbo frequency of up to 4.1 GHz, ensuring stable computing performance.

    • For information about OS compatibility, see AMD instance types and OS compatibility.

  • Storage:

  • Network:

  • Security: Supports the Trusted Computing (vTPM) feature.

The following table lists the specifications for the c9a instance family.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Packet rate (PPS)

Connections

NIC queues

ENIs

Private IPv4 addresses

IPv6 addresses

Disk IOPS

Disk bandwidth (Gbit/s)

ecs.c9a.large

2

4

2.5/Up to 15

Up to 1,200,000

Up to 500,000

2

3

6

6

Up to 110,000

2/Up to 15

ecs.c9a.xlarge

4

8

4/Up to 15

Up to 1,400,000

Up to 500,000

4

4

15

15

Up to 110,000

3/Up to 15

ecs.c9a.2xlarge

8

16

6/Up to 15

Up to 2,000,000

Up to 500,000

8

4

15

15

Up to 110,000

4/Up to 15

ecs.c9a.4xlarge

16

32

12/Up to 25

Up to 3,000,000

Up to 500,000

16

8

30

30

Up to 110,000

5/Up to 15

ecs.c9a.8xlarge

32

64

16/Up to 32

Up to 4,000,000

Up to 800,000

32

8

30

30

Up to 110,000

8/Up to 15

ecs.c9a.16xlarge

64

128

32/None

7,500,000

1,500,000

64

8

30

30

120,000

16/None

Compute-optimized instance family c9i

  • Introduction: Built on Alibaba Cloud's next-generation CIPU architecture and powered by Intel® Xeon® 6 processors with P-cores, these instances deliver stable computing performance, a powerful I/O engine, and chip-level security hardening.

  • Use cases: Machine learning inference applications, data analytics, batch computing, video encoding, frontend game servers, high-performance scientific and engineering applications, and web frontend servers.

  • Compute:

    • The vCPU-to-memory ratio is 1:2.

    • Processor: Intel® Xeon® Granite Rapids processors with a 3.2 GHz base frequency, an all-core turbo frequency of 3.6 GHz, and a maximum single-core turbo frequency of 3.9 GHz.

      Note

      The system may display different frequencies for this instance. The 3.9 GHz maximum single-core turbo frequency is a burst capability that depends on the overall CPU load of the physical server and is not covered by the Service Level Agreement (SLA).

    • For operating system compatibility, see Intel instance types and operating system compatibility.

  • Storage:

    • I/O-optimized instance.

    • These instances support the NVMe protocol. For more information, see Overview of the NVMe protocol.

    • Supported cloud disk types: ESSDs and ESSD AutoPL disks. For more information about cloud disks, see Block storage overview.

    • For small instance types, disk IOPS and disk bandwidth are burstable. The storage I/O performance of an instance corresponds to its instance type. A larger instance type provides higher storage I/O performance. For more information, see Storage I/O performance.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

c9i includes the following instance types and metric data:

Instance type

vCPUs

Memory (GiB)

Network bandwidth (baseline/burst) (Gbit/s)

Network PPS

Connections

NIC queues

ENIs

Private IPv4s per ENI

IPv6 addresses per ENI

Cloud disk IOPS (baseline/burst)

Cloud disk bandwidth (baseline/burst) (Gbit/s)

ecs.c9i.large

2

4

2.5/15

1,000,000

500,000

2

3

6

6

25,000/200,000

2/10

ecs.c9i.xlarge

4

8

4/15

1,200,000

500,000

4

4

15

15

50,000/200,000

2.5/10

ecs.c9i.2xlarge

8

16

6/15

1,600,000

500,000

8

4

15

15

60,000/200,000

4/10

ecs.c9i.3xlarge

12

24

10/15

2,400,000

500,000

12

8

15

15

80,000/200,000

5/10

ecs.c9i.4xlarge

16

32

12/25

3,000,000

500,000

16

8

30

30

100,000/200,000

6/10

ecs.c9i.6xlarge

24

48

15/25

4,500,000

600,000

24

8

30

30

120,000/200,000

7.5/10

ecs.c9i.8xlarge

32

64

20/32

6,000,000

800,000

32

8

30

30

200,000/300,000

10/12

ecs.c9i.12xlarge

48

96

25/32

9,000,000

1,600,000

48

8

30

30

240,000/320,000

12/15

ecs.c9i.16xlarge

64

128

28/36

12,000,000

2,000,000

64

8

30

30

300,000/400,000

16/24

ecs.c9i.24xlarge

96

192

32/48

18,000,000

3,000,000

64

15

30

30

350,000/600,000

20/28

ecs.c9i.32xlarge

128

256

36/50

20,000,000

4,000,000

64

15

30

30

400,000/650,000

24/28

ecs.c9i.48xlarge

192

384

64/None

24,000,000

6,000,000

64

15

50

50

500,000/800,000

32/None

Compute-optimized instance family c8a

  • Introduction: Built on Alibaba Cloud's next-generation CIPU architecture, these instances deliver stable computing power, a powerful I/O engine, and chip-level security hardening.

  • Use cases: Big data applications, web applications, AI training and inference, and audio and video transcoding.

  • Compute:

    • The CPU-to-memory ratio is 1:2.

    • Processor: AMD EPYC Genoa processors with a 2.7 GHz base frequency and a turbo frequency of up to 3.7 GHz deliver stable computing power.

    • Supports hyper-threading. Hyper-threading is enabled by default. To change this setting, see Change CPU options.

    • For information about operating system compatibility, see AMD instance type and OS compatibility.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For IPv6, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Delivers an ultra-high packet forwarding rate (PPS).

    • Smaller instance types support burstable network bandwidth.

    • Network performance scales with the instance type.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

c8a includes the following instance types and metrics:

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Base/Burst) (Gbit/s)

Packet forwarding rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4s/ENI

IPv6s/ENI

Cloud disk IOPS (Base/Burst)

Cloud disk bandwidth (Base/Burst) (Gbit/s)

ecs.c8a.large

2

4

1.5/Up to 12.5

900,000

Up to 250,000

2

3

6

6

20,000/Up to 110,000

1.5/Up to 10

ecs.c8a.xlarge

4

8

2.5/Up to 12.5

1,000,000

Up to 250,000

4

4

15

15

30,000/Up to 110,000

2/Up to 10

ecs.c8a.2xlarge

8

16

4/Up to 12.5

1,600,000

Up to 250,000

8

4

15

15

45,000/Up to 110,000

2.5/Up to 10

ecs.c8a.4xlarge

16

32

7/Up to 12.5

2,000,000

300,000

16

8

30

30

60,000/Up to 110,000

3.5/Up to 10

ecs.c8a.8xlarge

32

64

10/Up to 25

3,000,000

600,000

32

8

30

30

80,000/Up to 110,000

5/Up to 10

ecs.c8a.12xlarge

48

96

16/25

4,500,000

750,000

48

8

30

30

120,000/None

8/Up to 10

ecs.c8a.16xlarge

64

128

20/25

6,000,000

1,000,000

64

8

30

30

160,000/None

10/None

ecs.c8a.24xlarge

96

192

32/None

9,000,000

1,500,000

64

15

30

30

240,000/None

16/None

ecs.c8a.32xlarge

128

256

40/None

12,000,000

2,000,000

64

15

30

30

320,000/None

20/None

ecs.c8a.48xlarge

192

384

64/None

18,000,000

3,000,000

64

15

50

50

500,000/None

32/None

Note

For ecs.c8a.large and ecs.c8a.xlarge instances, you must enable jumbo frames to achieve the burst network bandwidth of 12.5 Gbit/s. For more information, see jumbo frames.

Compute-optimized instance family c8i

  • Introduction: Built on the next-generation Alibaba Cloud CIPU architecture, c8i instances provide stable computing performance, a powerful I/O engine, and chip-level security hardening.

  • Use cases: Ideal for machine learning inference, data analytics, batch computing, video encoding, game frontend servers, high-performance scientific and engineering applications, and web frontend servers.

  • Compute:

    • The vCPU-to-memory ratio is 1:2.

    • Processor: Intel®Xeon®Emerald Rapids or Intel®Xeon®Sapphire Rapids processors with a base frequency of at least 2.7 GHz and an all-core turbo frequency of 3.2 GHz.

      Note

      When you purchase an instance of this family, the system randomly assigns one of the preceding processors. You cannot manually select a processor.

    • Supports Hyper-Threading, enabled by default. For more information, see Change CPU options.

    • For information about operating system compatibility, see Intel instance types and operating system compatibility.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For more information about IPv6, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Network performance scales with the instance type.

    • Smaller instance types support burstable network bandwidth.

  • Security:

    • These instances support the vTPM feature. For more information, see Overview of trusted computing.

    • The vTPM feature uses TPM/TCM chips to create a trusted boot chain from the server to the instance for enhanced security.

    • Instances with 4 or more vCPUs support the Alibaba Cloud virtualized enclave feature to provide a confidential computing environment. For more information, see Build a confidential computing environment by using an enclave.

    • Uses Intel® Total Memory Encryption (TME) to encrypt runtime memory.

The following table describes the instance types and specifications of the c8i family.

Instance type

vCPUs

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Cloud disk baseline/burst IOPS

Cloud disk baseline/burst bandwidth (Gbit/s)

ecs.c8i.large

2

4

2.5/Up to 15

1,000,000

Up to 300,000

2

3

6

6

25,000/Up to 200,000

2/Up to 10

ecs.c8i.xlarge

4

8

4/Up to 15

1,200,000

Up to 300,000

4

4

15

15

50,000/Up to 200,000

2.5/Up to 10

ecs.c8i.2xlarge

8

16

6/Up to 15

1,600,000

Up to 300,000

8

4

15

15

60,000/Up to 200,000

4/Up to 10

ecs.c8i.3xlarge

12

24

10/Up to 15

2,400,000

Up to 300,000

12

8

15

15

80,000/Up to 200,000

5/Up to 10

ecs.c8i.4xlarge

16

32

12/Up to 25

3,000,000

350,000

16

8

30

30

100,000/Up to 200,000

6/Up to 10

ecs.c8i.6xlarge

24

48

15/Up to 25

4,500,000

500,000

24

8

30

30

120,000/Up to 200,000

7.5/Up to 10

ecs.c8i.8xlarge

32

64

20/Up to 25

6,000,000

800,000

32

8

30

30

200,000/None

10/None

ecs.c8i.12xlarge

48

96

25/None

9,000,000

1,000,000

48

8

30

30

300,000/None

12/None

ecs.c8i.16xlarge

64

128

32/None

12,000,000

1.6 million

64

8

30

30

360,000/None

20/None

ecs.c8i.24xlarge

96

192

50/None

18 million

2 million

64

15

30

30

500,000/None

24/None

ecs.c8i.48xlarge

192

512

100/Unlimited

30 million

4 million

64

15

50

50

1,000,000/None

48/None

Network-enhanced compute-optimized instance family c8ine

  • Introduction: Built on the new-generation Alibaba Cloud CIPU architecture, these instances deliver stable computing performance and a powerful I/O engine.

  • Use cases: Ideal for network-intensive workloads that require high forwarding and connection performance, such as network access layer gateways, traffic and data forwarding, and pre-processing middleware. As part of a cloud-native solution, these instances support large websites, e-commerce, and AI applications.

  • Compute:

    • The vCPU-to-memory ratio is 1:2.

    • Processor: Powered by Intel® Xeon® Emerald Rapids or Intel® Xeon® Sapphire Rapids processors with a base frequency of at least 2.7 GHz and an all-core turbo frequency of 3.2 GHz.

    • Supports Hyper-Threading, which is enabled by default. For more information, see Change CPU options.

  • Storage:

    • These are I/O optimized instances.

    • These instances support the NVMe protocol. For more information, see Overview of the NVMe protocol.

    • Supported cloud disk types: ESSDs, ESSD AutoPL disks, and Regional ESSDs. For more information about cloud disks, see Block storage overview.

    • Smaller instance types offer burstable disk IOPS and disk bandwidth.

  • Network:

    • Supports IPv4 and IPv6. For more information about IPv6 communication, see IPv6 communication.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Network performance scales with the instance type.

  • Security: Supports trusted computing (vTPM). For more information, see Trusted computing overview.

The following table lists the instance types and specifications for the c8ine instance family.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Connections

Multi-queue

ENIs

Private IPv4 addresses

IPv6 addresses

ESSD multi-queue

Disk IOPS

Disk bandwidth (Gbit/s)

ecs.c8ine.large

2

4

4/24

600,000

2

3

10

10

1

20,000/80,000

2/8

ecs.c8ine.xlarge

4

8

7/28

1,200,000

4

4

15

15

1

40,000/80,000

2.5/8

ecs.c8ine.2xlarge

8

16

12/35

2,000,000

8

6

15

15

2

50,000/80,000

4/8

ecs.c8ine.4xlarge

16

32

23/44

3,500,000

16

8

30

30

2

80,000/100,000

6/10

ecs.c8ine.8xlarge

32

64

44/N/A

7,000,000

32

8

30

30

4

100,000/N/A

10/N/A

Compute-balanced enhanced instance family c8ae

  • Introduction: Built on the Alibaba Cloud-developed CIPU architecture, these instances provide stable computing performance, a powerful I/O engine, and chip-level security hardening.

  • Use cases:

    • AI workloads, such as deep learning, training, and AI inference.

    • High-performance computing (HPC) workloads.

    • Large and medium-sized database systems, caches, and search clusters.

    • Massively multiplayer online (MMO) game servers.

    • Other general-purpose enterprise applications with high performance requirements.

  • Compute:

    • The CPU-to-memory ratio is 1:2.

    • The AMD EPYC Genoa processor provides stable computing performance with a base frequency of 3.4 GHz and a single-core turbo frequency of up to 3.75 GHz.

    • Hyper-threading is enabled by default. To change this setting, see Change CPU options.

    • For operating system compatibility, see AMD instance type and OS compatibility.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For more information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Delivers an ultra-high packet forwarding rate (PPS).

    • Smaller instance types support burstable network bandwidth.

    • Network performance scales with the instance type.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

c8ae includes the instance types and metrics listed in the following table:

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

PPS

vTPM

Connections

NIC queues

ENIs

Private IPv4/ENI

IPv6/ENI

Cloud disk IOPS

Disk bandwidth (Gbit/s)

ecs.c8ae.large

2

4

3/Up to 15

1,000,000

Yes

Up to 300,000

2

3

6

6

30,000/Up to 200,000

2/Up to 10

ecs.c8ae.xlarge

4

8

4/Up to 15

1,200,000

Yes

Up to 300,000

4

4

15

15

50,000/Up to 200,000

2.5/Up to 10

ecs.c8ae.2xlarge

8

16

6/Up to 15

1,600,000

Yes

Up to 300,000

8

4

15

15

60,000/Up to 200,000

3/Up to 10

ecs.c8ae.4xlarge

16

32

12/Up to 25

3,000,000

Yes

500,000

16

8

30

30

100,000/Up to 200,000

6/Up to 10

ecs.c8ae.8xlarge

32

64

20/Up to 25

6,000,000

Yes

1,000,000

32

8

30

30

200,000/None

10/None

ecs.c8ae.16xlarge

64

128

32/None

9,000,000

Yes

1,500,000

64

8

30

30

250,000/None

16/None

ecs.c8ae.32xlarge

128

256

64/None

18,000,000

Yes

3,000,000

64

15

30

30

500,000/None

32/None

Note

To achieve the 15 Gbit/s burst network bandwidth on ecs.c8ae.large and ecs.c8ae.xlarge instances, you must enable jumbo frames. For more information, see Jumbo Frames.

Compute-optimized instance family c7a

  • Introduction: Built on the third-generation SHENLONG architecture, c7a instances deliver stable and predictable ultra-high performance. They use chip-level fast path acceleration to achieve order-of-magnitude improvements in storage performance, network performance, and compute stability.

  • Use cases:

    • Video encoding and decoding.

    • High packet forwarding rate (PPS) workloads, such as live video commenting and telecom service forwarding.

    • Web frontend servers.

    • Massively multiplayer online (MMO) game frontends.

    • Development and testing, such as DevOps.

    • Data analytics and batch computing.

    • High-performance scientific and engineering applications.

    • Enterprise-grade applications of all types and sizes.

  • Compute:

    • The vCPU-to-memory ratio is 1:2.

    • Processor: AMD EPYC Milan processors with a 2.55 GHz base frequency and a single-core turbo frequency of up to 3.5 GHz deliver stable compute performance.

    • Hyper-Threading is supported and enabled by default. For more information, see Change CPU options.

    • For information about operating system compatibility, see AMD instance type and OS compatibility.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For more information, see IPv6 communication.

    • Provides an ultra-high packet forwarding rate (PPS).

    • Smaller instance types support burstable network bandwidth.

    • Network performance increases with instance size.

c7a includes the instance types and metrics listed in the following table:

Instance type

vCPUs

Memory (GiB)

Network bandwidth (baseline/burst) (Gbit/s)

Packet forwarding rate (PPS)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Cloud disk IOPS (baseline/burst)

Cloud disk bandwidth (baseline/burst) (Gbit/s)

ecs.c7a.large

2

4

1/10

900,000

250,000

2

3

6

6

12,500/110,000

1/6

ecs.c7a.xlarge

4

8

1.5/10

1,000,000

250,000

4

4

15

15

20,000/110,000

1.5/6

ecs.c7a.2xlarge

8

16

2.5/10

1,600,000

250,000

8

4

15

15

30,000/110,000

2/6

ecs.c7a.4xlarge

16

32

5/10

2,000,000

300,000

8

8

30

30

60,000/110,000

3/6

ecs.c7a.8xlarge

32

64

8/10

3,000,000

600,000

16

7

30

30

75,000/110,000

4/6

ecs.c7a-nps1.8xlarge

32

64

8/10

3,000,000

600,000

16

7

30

30

75,000/110,000

4/6

ecs.c7a.16xlarge

64

128

16/N/A

6,000,000

1,000,000

32

7

30

30

150,000/N/A

8/N/A

ecs.c7a-nps1.16xlarge

64

128

16/N/A

3,000,000

1,000,000

32

7

30

30

150,000/N/A

8/N/A

ecs.c7a.32xlarge

128

256

32/N/A

12,000,000

2,000,000

32

15

30

30

300,000/N/A

16/N/A

Note

The kernels of Ubuntu 16 and Debian 9 do not support AMD EPYC Milan processors. Do not use Ubuntu 16 or Debian 9 images to create c7a instances. Otherwise, the instances fail to start.

Compute-optimized instance family c7

  • Introduction: Built on the third-generation Shenlong architecture, c7 instances deliver stable and predictable ultra-high performance. They use chip-level fast-path acceleration to deliver dramatic improvements in storage, network performance, and computing stability.

  • Use cases:

    • Scenarios with high packet forwarding rates (PPS), such as real-time video comment streaming and telecom service forwarding.

    • Front-end servers for massively multiplayer online (MMO) games.

    • Web front-end servers.

    • Data analytics, batch computing, and video encoding.

    • High-performance scientific and engineering applications.

    • Workloads that require trusted computing.

    • Enterprise-grade applications of all types and sizes.

    • Blockchain applications.

  • Compute:

    • vCPU-to-memory ratio of 1:2.

    • Processor: Features 3rd Generation Intel® Xeon® Scalable (Ice Lake) processors with a base frequency of 2.7 GHz and an all-core turbo frequency of 3.5 GHz.

    • Hyper-Threading is supported and enabled by default. For more information, see Configure CPU options.

  • Storage:

  • Network:

    • Supports both IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Provides an ultra-high packet forwarding rate (PPS).

    • Smaller instance types support burstable network bandwidth.

    • Network performance scales with the instance type.

  • Security:

The following table lists the specifications of the c7 instance types.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

PPS

vTPM support

Connections

Multi-queue

ENIs

Private IPv4 addresses/ENI

IPv6 addresses/ENI

Max data disks

Cloud disk IOPS

Cloud disk bandwidth (Gbit/s)

ecs.c7.large

2

4

2/Up to 12.5

1,100,000

Yes

Up to 500,000

2

3

6

6

8

20,000/Up to 160,000

1.5/Up to 10

ecs.c7.xlarge

4

8

3/Up to 12.5

1,100,000

Yes

Up to 500,000

4

4

15

15

8

40,000/Up to 160,000

2/Up to 10

ecs.c7.2xlarge

8

16

5/Up to 15

1,600,000

Yes

Up to 500,000

8

4

15

15

16

50,000/Up to 160,000

3/Up to 10

ecs.c7.3xlarge

12

24

8/Up to 15

2,400,000

Yes

Up to 500,000

8

8

15

15

16

70,000/Up to 160,000

4/Up to 10

ecs.c7.4xlarge

16

32

10/Up to 25

3,000,000

Yes

500,000

8

8

30

30

16

80,000/Up to 160,000

5/Up to 10

ecs.c7.6xlarge

24

48

12/Up to 25

4,500,000

Yes

550,000

12

8

30

30

16

110,000/Up to 160,000

6/Up to 10

ecs.c7.8xlarge

32

64

16/Up to 32

6,000,000

Yes

600,000

16

8

30

30

24

160,000/None

10/None

ecs.c7.16xlarge

64

128

32/None

12,000,000

Yes

1,200,000

32

8

30

30

32

360,000/None

16/None

ecs.c7.32xlarge

128

256

64/None

24,000,000

Yes

2,400,000

32

15

30

30

32

600,000/None

32/None

Compute-optimized instance family c6

  • Introduction: Built on the Shenlong architecture, c6 instances offload many virtualization features to dedicated hardware, reducing virtualization overhead and delivering stable, predictable, and ultra-high performance.

  • Use cases:

    • High packet forwarding rate (PPS) workloads, such as live video commenting and telecom service forwarding.

    • Web frontend servers.

    • Frontend servers for massively multiplayer online (MMO) games.

    • Data analytics, batch computing, and video encoding

    • High-performance scientific and engineering applications.

  • Compute:

    • The vCPU-to-memory ratio is 1:2.

    • Processor: Powered by 2.5 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processors with a 3.2 GHz turbo frequency, delivering stable computing performance.

    • Hyper-Threading is supported and enabled by default. For more information, see Change CPU options.

    Note

    This instance family does not support FreeBSD 13.2 or earlier versions due to compatibility issues. Use FreeBSD 13.3 or later.

  • Storage:

    • I/O optimized instance.

    • Supported cloud disk types: ESSDs, ESSD AutoPL disks, standard SSDs, and ultra disks. For more information, see Block storage overview.

      Note

      The maximum performance of disks varies based on the instance families. A single instance of this instance family can deliver up to 200,000 IOPS.

    • Storage I/O performance scales with the instance type. For more information, see Storage I/O performance.

  • Network:

    • Supports IPv4 and IPv6. For more information about IPv6 communication, see IPv6 communication.

    • Ultra-high packet forwarding rate (PPS).

    • Network performance scales with the instance type.

The c6 instance types and their specifications are listed in the following table.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (baseline/burst, Gbit/s)

Packet forwarding rate (PPS)

Connections

Multi-queue

Elastic network interfaces

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Cloud disk baseline IOPS

Cloud disk baseline bandwidth (Gbit/s)

ecs.c6.large

2

4

1/up to 3

300,000

up to 250,000

2

2

6

1

10,000

1

ecs.c6.xlarge

4

8

1.5/up to 5

500,000

up to 250,000

4

3

10

1

20,000

1.5

ecs.c6.2xlarge

8

16

2.5/up to 8

800,000

up to 250,000

8

4

10

1

25,000

2

ecs.c6.3xlarge

12

24

4/up to 10

900,000

up to 250,000

8

6

10

1

30,000

2.5

ecs.c6.4xlarge

16

32

5/up to 10

1,000,000

300,000

8

8

20

1

40,000

3

ecs.c6.6xlarge

24

48

7.5/up to 10

1,500,000

450,000

12

8

20

1

50,000

4

ecs.c6.8xlarge

32

64

10/None

2,000,000

600,000

16

8

20

1

60,000

5

ecs.c6.13xlarge

52

96

12.5/None

3,000,000

900,000

32

7

20

1

100,000

8

ecs.c6.26xlarge

104

192

25/None

6,000,000

1,800,000

32

15

20

1

200,000

16

Compute-optimized instance family c6a

  • Introduction: Built on the Shenlong architecture, c6a instances offload many virtualization features to dedicated hardware. This approach reduces virtualization overhead and delivers stable, predictable, and ultra-high performance.

  • Use cases: Video encoding and decoding, high packet rate (PPS) workloads, web frontend servers, frontend servers for massively multiplayer online (MMO) games, and development and testing (DevOps).

  • Compute:

    • The vCPU-to-memory ratio is 1:2.

    • Processor: Powered by 2.6 GHz AMD EPYC ROME processors with a 3.3 GHz turbo frequency for stable compute performance.

    • Hyper-Threading is supported and enabled by default. For more information, see Change CPU options.

    • For information about operating system compatibility, see AMD instance type and OS compatibility.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances deliver an ultra-high packet rate (PPS).

    • Network performance scales with the instance type.

The following table describes the instance types and specifications of the c6a family.

Instance type

vCPUs

Memory (GiB)

Bandwidth (base/burst) (Gbit/s)

Packet rate (PPS)

Connections

NIC queues

ENIs

Private IPv4 addresses/ENI

IPv6 addresses/ENI

Base disk IOPS

Base disk bandwidth (Gbit/s)

ecs.c6a.large

2

4

1/10

900,000

Up to 250,000

2

2

6

1

12,500

1

ecs.c6a.xlarge

4

8

1.5/10

1,000,000

Up to 250,000

4

3

15

1

20,000

1.5

ecs.c6a.2xlarge

8

16

2.5/10

1,600,000

Up to 250,000

8

4

15

1

30,000

2

ecs.c6a.4xlarge

16

32

5/10

2,000,000

300,000

8

8

30

1

60,000

3.1

ecs.c6a.8xlarge

32

64

8/10

3,000,000

600,000

16

7

30

1

75,000

4.1

ecs.c6a.16xlarge

64

128

16/None

6,000,000

1,000,000

32

8

30

1

150,000

8.2

ecs.c6a.32xlarge

128

256

32/None

12,000,000

2,000,000

32

15

30

1

300,000

16.4

Compute-balanced instance family c6e

  • Introduction: Built on the third-generation SHENLONG architecture, c6e instances offload many virtualization features to dedicated hardware. This design reduces virtualization overhead and delivers stable, predictable, and ultra-high performance. These instances also use chip-level fast-path acceleration to achieve order-of-magnitude improvements in storage performance, network performance, and compute stability.

  • Use cases:

    • High forwarding rate (PPS) workloads, such as live video commenting and telecom traffic forwarding.

    • Web frontend servers.

    • Massively multiplayer online (MMO) game frontends.

    • Data analytics, batch computing, and video encoding.

    • High-performance scientific and engineering applications.

  • Compute:

    • The vCPU-to-memory ratio is approximately 1:2.

    • Processor: Intel® Xeon® Platinum 8269CY (Cascade Lake) with a 2.5 GHz base frequency and a 3.2 GHz turbo frequency, delivering stable compute performance.

    • Supports Hyper-Threading. This feature is enabled by default. For more information, see Change CPU options.

    Note

    This instance family does not support FreeBSD 13.2 or earlier versions due to compatibility issues. You must use FreeBSD 13.3 or a later version.

  • Storage:

  • Network:

    • Supports both IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Offers an ultra-high forwarding rate (PPS).

      Note

      Network performance varies based on the instance families. For higher concurrent connection and network packet forwarding capabilities, we recommend that you use the g7ne instance family.

    • Network performance scales with the instance type. Larger instance types deliver better network performance.

This table lists the specifications for the c6e instance family.

Instance type

vCPUs

Memory (GiB)

Bandwidth (Gbit/s)

Forwarding rate (PPS)

Connections

NIC queues

ENIs

Private IPv4s per ENI

IPv6s per ENI

Disk baseline IOPS

Disk bandwidth (Gbit/s)

ecs.c6e.large

2

4

1.2/10

900,000

Up to 250,000

2

3

6

1

20,000

1

ecs.c6e.xlarge

4

8

2/10

1,000,000

Up to 250,000

4

4

15

1

40,000

1.5

ecs.c6e.2xlarge

8

16

3/10

1,600,000

Up to 250,000

8

4

15

1

50,000

2

ecs.c6e.4xlarge

16

32

6/10

3,000,000

300,000

8

8

30

1

80,000

3

ecs.c6e.8xlarge

32

64

10/N/A

6,000,000

600,000

16

8

30

1

150,000

5

ecs.c6e.13xlarge

52

96

16/N/A

9,000,000

1,000,000

32

7

30

1

240,000

8

ecs.c6e.26xlarge

104

192

32/N/A

24,000,000

1,800,000

32

15

30

1

480,000

16

Compute-optimized instance family c5

  • Use cases:

    • High packet rate (PPS) workloads, such as live video commenting and telecom service forwarding.

    • web frontend servers.

    • Massively multiplayer online (MMO) game frontends.

    • Data analytics, batch computing, and video encoding.

    • High-performance scientific and engineering applications.

  • Compute:

    • The vCPU-to-memory ratio is 1:2.

    • Processor: 2.5 GHz Intel® Xeon® Platinum 8163 (Skylake) or 8269CY (Cascade Lake).

      Note

      Instances of this family may be deployed on different server platforms. If your workload requires a consistent server platform, use c6, c6e, or c7 instances.

  • Storage:

    • I/O-optimized instance.

    • Supported disk types: ESSDs, ESSD AutoPL disks, Standard SSDs, and Ultra Disks. For more information, see block storage.

      Note

      The maximum performance of disks varies based on the instance families. A single instance of this instance family can deliver up to 200,000 IOPS.

  • Network:

    • Supports IPv4 and IPv6. For more information about IPv6 communication, see IPv6 communication.

    • Provides high network PPS performance.

    • Network performance scales with the instance type.

The c5 family includes the following instance types and specifications:

Instance type

vCPUs

Memory (GiB)

Baseline bandwidth (Gbit/s)

Packet rate (PPS)

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

ecs.c5.large

2

4

1

300,000

2

2

6

1

ecs.c5.xlarge

4

8

1.5

500,000

2

3

10

1

ecs.c5.2xlarge

8

16

2.5

800,000

4

4

10

1

ecs.c5.3xlarge

12

24

4

900,000

4

6

10

1

ecs.c5.4xlarge

16

32

5

1,000,000

4

8

20

1

ecs.c5.6xlarge

24

48

7.5

1,500,000

6

8

20

1

ecs.c5.8xlarge

32

64

10

2,000,000

8

8

20

1

ecs.c5.16xlarge

64

128

20

4,000,000

16

8

20

1

Compute-intensive instance family ic5

  • Use cases:

    • Web frontend servers.

    • Data analytics, batch computing, and video encoding.

    • High packet forwarding rate (PPS) workloads, such as live video commenting and telecom service forwarding.

    • Frontend servers for massively multiplayer online (MMO) games.

  • Compute:

    • The vCPU-to-memory ratio is 1:1.

    • Processor: Intel® Xeon® Platinum 8163 (Skylake) or 8269CY (Cascade Lake) with a base frequency of 2.5 GHz and an all-core turbo frequency of 2.7 GHz, delivering stable compute performance.

  • Storage:

  • Network:

    • Supports only IPv4.

    • Delivers an ultra-high packet forwarding rate (PPS).

    • Network performance scales with the instance type.

The following table lists the instance types and specifications for the ic5 family.

Instance type

vCPU

Memory (GiB)

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (PPS)

NIC queues

ENIs

Private IPv4s per ENI

ecs.ic5.large

2

2

1

300,000

2

2

6

ecs.ic5.xlarge

4

4

1.5

500,000

2

3

10

ecs.ic5.2xlarge

8

8

2.5

800,000

2

4

10

ecs.ic5.3xlarge

12

12

4

900,000

4

6

10

ecs.ic5.4xlarge

16

16

5

1,000,000

4

8

20

ecs.ic5.6xlarge

24

24

7.5

1,500,000

6

8

20

ecs.ic5.8xlarge

32

32

10

2,000,000

8

8

20

ecs.ic5.16xlarge

64

64

20

3,000,000

16

8

20

Compute and network-enhanced instance family sn1ne

  • Use cases:

    • High packet rate (PPS) workloads, such as live video commenting and telecom service forwarding.

    • Web frontend servers.

    • Massively multiplayer online (MMO) game frontends.

    • Data analytics, batch computing, and video encoding.

    • High-performance scientific and engineering applications.

  • Compute:

    • The vCPU-to-memory ratio is 1:2.

    • Processor: Powered by 2.5 GHz Intel ® Xeon ® E5-2682 v4 (Broadwell), Platinum 8163 (Skylake), or 8269CY (Cascade Lake) processors.

      Note

      Instances in this family may be deployed on different server platforms. If your application requires a consistent server platform, we recommend that you use c6, c6e, or c7 instances.

  • Storage:

    • I/O-optimized instance.

    • Supported cloud disk types: standard SSDs and ultra disks. For more information, see block storage.

  • Network:

    • Supports IPv4 and IPv6. For more information, see IPv6 communication.

    • Delivers an ultra-high packet rate (PPS).

    • Network performance scales with the instance type.

The sn1ne instance family includes the following instance types:

Instance type

vCPUs

Memory (GiB)

Baseline bandwidth (Gbit/s)

Packet rate (PPS)

NIC queues

ENIs

Private IPv4 addresses

IPv6 addresses

ecs.sn1ne.large

2

4

1

300,000

2

2

6

1

ecs.sn1ne.xlarge

4

8

1.5

500,000

2

3

10

1

ecs.sn1ne.2xlarge

8

16

2

1,000,000

4

4

10

1

ecs.sn1ne.3xlarge

12

24

2.5

1,300,000

4

6

10

1

ecs.sn1ne.4xlarge

16

32

3

1,600,000

4

8

20

1

ecs.sn1ne.6xlarge

24

48

4.5

2,000,000

6

8

20

1

ecs.sn1ne.8xlarge

32

64

6

2,500,000

8

8

20

1

Memory-optimized instance family r9ae

  • Introduction: This instance family is built on the new Alibaba Cloud Infrastructure Processing Unit (CIPU) architecture and is powered by the latest AMD EPYC Turin processors. Its physical core design ensures stable computing power, a powerful I/O engine, and chip-level security hardening.

  • Use cases: Big data analytics (such as Spark, Flink, and Elasticsearch), search, recommendation, and advertising (ps-worker), core transaction systems, audio and video transcoding, AI training and inference, and general-purpose enterprise applications (such as Java).

  • Compute:

    • The vCPU-to-memory ratio is 1:8.

    • Processor: AMD EPYC Turin processor with a turbo frequency of up to 3.7 GHz. The physical core design ensures stable computing performance.

    • For information about operating system compatibility, see AMD instance and operating system compatibility.

  • Storage:

  • Network:

  • Security:

The following table lists the instance types and specifications of the r9ae instance family.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Packet rate (PPS)

Connections

NIC queues

ENIs

Private IPv4 addresses

IPv6 addresses

Baseline disk IOPS

Disk bandwidth (Gbit/s)

ecs.r9ae.large

2

16

2.5/Up to 25

Up to 1,500,000

Up to 500,000

2

3

6

6

Up to 200,000

2.5/Up to 20

ecs.r9ae.xlarge

4

32

4/Up to 25

Up to 1,600,000

Up to 500,000

4

4

15

15

Up to 200,000

3/Up to 20

ecs.r9ae.2xlarge

8

64

6/Up to 25

Up to 2,500,000

Up to 500,000

8

4

15

15

Up to 200,000

4/Up to 20

ecs.r9ae.4xlarge

16

128

10/Up to 25

Up to 3,200,000

Up to 500,000

16

8

30

30

Up to 200,000

5.5/Up to 20

ecs.r9ae.8xlarge

32

256

16/Up to 25

Up to 5,000,000

Up to 1,000,000

32

8

30

30

Up to 200,000

8/Up to 20

ecs.r9ae.12xlarge

48

384

25/None

7,500,000

1,500,000

48

8

30

30

150,000

13/None

ecs.r9ae.16xlarge

64

512

32/None

10,000,000

2,000,000

64

8

30

30

200,000

16/None

ecs.r9ae.24xlarge

96

768

50/None

15,000,000

3,000,000

64

15

30

30

300,000

25/None

ecs.r9ae.32xlarge

128

1024

64/None

20,000,000

4,000,000

64

15

30

30

400,000

32/None

ecs.r9ae.48xlarge

192

1536

100/None

30,000,000

6,000,000

64

15

50

50

600,000

50/None

R9a memory-optimized instance family

  • Introduction: This instance family is built on the new Alibaba Cloud CIPU architecture and is powered by the latest AMD EPYC Turin processors. It provides stable computing power, a powerful I/O engine, and chip-level security hardening.

  • Use cases: Medium- to large-scale database systems, game servers, quantitative finance, blockchain, website and application servers, and other general-purpose enterprise applications.

  • Compute:

  • Storage:

  • Network:

  • Security: Supports vTPM.

The following table describes the instance types and specifications of the r9a instance family.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Packet rate (PPS)

Connections

NIC queues

ENIs

Private IPv4 addresses

IPv6 addresses

Baseline IOPS

Disk bandwidth (Gbit/s)

ecs.r9a.large

2

16

2.5/Up to 15

Up to 1.2 million

Up to 500,000

2

3

6

6

Up to 110,000

2/Up to 15

ecs.r9a.xlarge

4

32

4/Up to 15

Up to 1.4 million

Up to 500,000

4

4

15

15

Up to 110,000

3/Up to 15

ecs.r9a.2xlarge

8

64

6/Up to 15

Up to 2 million

Up to 500,000

8

4

15

15

Up to 110,000

4/Up to 15

ecs.r9a.4xlarge

16

128

12/Up to 25

Up to 3 million

Up to 500,000

16

8

30

30

Up to 110,000

5/Up to 15

ecs.r9a.8xlarge

32

256

16/Up to 32

Up to 4 million

Up to 800,000

32

8

30

30

Up to 110,000

8/Up to 15

Memory-optimized instance family r9i

  • Introduction: This instance family is built on the Alibaba Cloud new-generation CIPU architecture and is powered by Intel® Xeon® 6 processors with P-cores. It delivers stable computing power, an enhanced I/O engine, and chip-level security hardening.

  • Use cases: Data analytics and mining, Hadoop and Spark clusters, other memory-intensive enterprise applications, distributed in-memory caches such as Redis, and websites and application servers.

  • Compute:

    • The vCPU-to-memory ratio is 1:8.

    • Processor: Intel® Xeon® Granite Rapids processors with a base frequency of 3.2 GHz, an all-core turbo frequency of 3.6 GHz, and a maximum single-core turbo frequency of 3.9 GHz.

      Note

      This instance may show different frequencies in the operating system. The maximum single-core turbo frequency of 3.9 GHz represents burst performance. This burst performance depends on the overall CPU load on the physical host and is not guaranteed by a service-level agreement (SLA).

    • For information about OS compatibility, see OS compatibility for Intel instance types.

  • Storage:

    • This instance family consists of I/O-optimized instances.

    • These instances support the NVMe protocol. For more information, see Overview of the NVMe protocol.

    • Supported cloud disk types: ESSDs, ESSD AutoPL disks, and Regional ESSDs. For more information about cloud disks, see Block Storage overview.

    • For small instance types, disk IOPS and disk bandwidth are burstable. The storage I/O performance of an instance corresponds to its instance type. A larger instance type provides higher storage I/O performance. For more information, see Storage I/O performance.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

The following table lists the instance types in the r9i family and their specifications.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

NIC queues

ENIs

Private IPv4s per ENI

IPv6 addresses per ENI

Cloud disk IOPS

Cloud disk bandwidth (Gbit/s)

ecs.r9i.large

2

16

2.5/Up to 15

1,000,000

Up to 500,000

2

3

6

6

25,000/Up to 200,000

2/Up to 10

ecs.r9i.xlarge

4

32

4/Up to 15

1,200,000

Up to 500,000

4

4

15

15

50,000/Up to 200,000

2.5/Up to 10

ecs.r9i.2xlarge

8

64

6/Up to 15

1,600,000

Up to 500,000

8

4

15

15

60,000/Up to 200,000

4/Up to 10

ecs.r9i.3xlarge

12

96

10/Up to 15

2,400,000

Up to 500,000

12

8

15

15

80,000/Up to 200,000

5/Up to 10

ecs.r9i.4xlarge

16

128

12/Up to 25

3,000,000

500,000

16

8

30

30

100,000/Up to 200,000

6/Up to 10

ecs.r9i.6xlarge

24

192

15/Up to 25

4,500,000

600,000

24

8

30

30

120,000/Up to 200,000

7.5/Up to 10

ecs.r9i.8xlarge

32

256

20/Up to 32

6,000,000

800,000

32

8

30

30

200,000/Up to 300,000

10/Up to 12

ecs.r9i.12xlarge

48

384

25/Up to 32

9,000,000

1,600,000

48

8

30

30

240,000/Up to 320,000

12/Up to 15

ecs.r9i.16xlarge

64

512

28/Up to 36

12,000,000

2,000,000

64

8

30

30

300,000/Up to 400,000

16/Up to 24

ecs.r9i.24xlarge

96

768

32/Up to 48

18,000,000

3,000,000

64

15

30

30

350,000/Up to 600,000

20/Up to 28

ecs.r9i.32xlarge

128

1024

36/Up to 50

20,000,000

4,000,000

64

15

30

30

400,000/Up to 650,000

24/Up to 28

ecs.r9i.48xlarge

192

1536

64/N/A

24,000,000

6,000,000

64

15

50

50

500,000/Up to 800,000

32/N/A

r8a memory-optimized instance family

  • Introduction: Built on the new Alibaba Cloud CIPU architecture, this instance family provides stable computing power, a powerful I/O engine, and chip-level security hardening.

  • Use cases:

    • General-purpose, enterprise-level applications with high memory requirements, such as Java applications.

    • Various in-memory database applications, such as Redis and Memcache.

    • Big data applications, such as Kafka and Elasticsearch.

    • Audio and video transcoding applications.

  • Compute:

    • The vCPU-to-memory ratio is 1:8.

    • Processor: AMD EPYC™ Genoa processors with a base frequency of 2.7 GHz and a turbo frequency of up to 3.7 GHz.

    • Supports Hyper-Threading. Hyper-Threading is enabled by default. For more information, see Change CPU options.

    • For operating system compatibility, see AMD instance and operating system compatibility.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Supports Elastic RDMA Interface (ERI). To use ERI, see Enable eRDMA on an enterprise-grade instance.

    • Supports Jumbo Frames. For more information, see Jumbo Frames.

    • Provides a high packet forwarding rate (pps).

    • Smaller instance types feature burstable network bandwidth.

    • Network performance scales with the instance type.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

The following table describes the specifications of the r8a instance family.

Instance type

vCPUs

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Cloud disk baseline/burst IOPS

Cloud disk baseline/burst bandwidth (Gbit/s)

ecs.r8a.large

2

16

1.5/Up to 12.5

900,000

Up to 250,000

2

3

6

6

20,000/Up to 110,000

1.5/Up to 10

ecs.r8a.xlarge

4

32

2.5/Up to 12.5

1,000,000

Up to 250,000

4

4

15

15

30,000/Up to 110,000

2/Up to 10

ecs.r8a.2xlarge

8

64

4/Up to 12.5

1,600,000

Up to 250,000

8

4

15

15

45,000/Up to 110,000

2.5/Up to 10

ecs.r8a.4xlarge

16

128

7/Up to 12.5

2,000,000

300,000

16

8

30

30

60,000/Up to 110,000

3.5/Up to 10

ecs.r8a.8xlarge

32

256

10/Up to 25

3,000,000

600,000

32

8

30

30

80,000/Up to 110,000

5/Up to 10

ecs.r8a.12xlarge

48

384

16/25

4,500,000

750,000

48

8

30

30

120,000/None

8/Up to 10

ecs.r8a.16xlarge

64

512

20/25

6,000,000

1,000,000

64

8

30

30

160,000/None

10/None

ecs.r8a.24xlarge

96

768

32/None

9,000,000

1,500,000

64

15

30

30

240,000/None

16/None

ecs.r8a.32xlarge

128

1024

40/None

12,000,000

2,000,000

64

15

30

30

320,000/None

20/None

ecs.r8a.48xlarge

192

1536

64/None

18,000,000

3,000,000

64

15

50

50

500,000/None

32/None

Note
  • For ecs.r8a.large and ecs.r8a.xlarge instances, you must enable Jumbo Frames to achieve the burst network bandwidth of 12.5 Gbit/s. For more information, see Jumbo Frames.

  • To use the ecs.r8a.48xlarge instance type, submit a ticket.

Memory-optimized instance family r8i

  • Introduction: This instance family is built on the new Alibaba Cloud CIPU architecture and provides stable computing power, a powerful I/O engine, and chip-level security.

  • Use cases:

    • Data analytics and mining.

    • Hadoop and Spark clusters, and other enterprise applications with high memory requirements.

    • Distributed in-memory caches, such as Redis.

    • Websites and application servers.

    • Massively multiplayer online (MMO) game servers.

  • Compute:

    • The vCPU-to-memory ratio is 1:8.

    • Processor: Powered by Intel® Xeon® Emerald Rapids or Intel® Xeon® Sapphire Rapids processors with a base frequency of at least 2.7 GHz and an all-core turbo frequency of 3.2 GHz, ensuring stable computing performance.

    • Hyper-Threading is supported and enabled by default. For more information, see Change CPU options.

    • For information about operating system compatibility, see Intel instance types and operating system compatibility.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Supports ERI (elastic rdma interface). For information about how to use ERI, see Enable eRDMA on enterprise-level instances.

    • Supports jumbo frames. For more information, see Jumbo Frames.

    • Smaller instance types offer burstable network bandwidth.

    • Network performance scales with the instance type.

  • Security:

The following table lists the specifications for the r8i instance family.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Packet rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk IOPS

Disk bandwidth (Gbit/s)

ecs.r8i.large

2

16

2.5/Up to 15

1,000,000

Up to 300,000

2

3

6

6

25,000/Up to 200,000

2/Up to 10

ecs.r8i.xlarge

4

32

4/Up to 15

1,200,000

Up to 300,000

4

4

15

15

50,000/Up to 200,000

2.5/Up to 10

ecs.r8i.2xlarge

8

64

6/Up to 15

1,600,000

Up to 300,000

8

4

15

15

60,000/Up to 200,000

4/Up to 10

ecs.r8i.3xlarge

12

96

10/Up to 15

2,400,000

Up to 300,000

12

8

15

15

80,000/Up to 200,000

5/Up to 10

ecs.r8i.4xlarge

16

128

12/Up to 25

3,000,000

350,000

16

8

30

30

100,000/Up to 200,000

6/Up to 10

ecs.r8i.6xlarge

24

192

15/Up to 25

4,500,000

500,000

24

8

30

30

120,000/Up to 200,000

7.5/Up to 10

ecs.r8i.8xlarge

32

256

20/Up to 25

6,000,000

800,000

32

8

30

30

200,000/N/A

10/N/A

ecs.r8i.12xlarge

48

384

25/N/A

9,000,000

1,000,000

48

8

30

30

300,000/N/A

12/N/A

ecs.r8i.16xlarge

64

512

32/N/A

12,000,000

1,600,000

64

8

30

30

360,000/N/A

20/N/A

ecs.r8i.32xlarge

128

1024

64/N/A

24,000,000

3,000,000

64

15

30

30

700,000/N/A

40/N/A

Note

To use the ecs.r8i.16xlarge and ecs.r8i.32xlarge instance types, submit a ticket.

r8ae: Memory-balanced enhanced instance family

  • Introduction: This instance family is built on the Alibaba Cloud CIPU architecture. It delivers stable computing power, a powerful I/O engine, and chip-level security hardening.

  • Scenarios:

    • AI workloads, such as deep learning, training, and AI inference.

    • High-performance computing (HPC) and other scientific computing scenarios.

    • Medium- to large-sized database systems, caches, and search clusters.

    • Large online game servers.

    • Other general-purpose enterprise applications with high performance requirements.

  • Compute:

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For IPv6 communication, see IPv6 communication.

    • Supports Elastic RDMA Interface (ERI). See Enable eRDMA on enterprise-level instances for instructions.

    • Supports jumbo frames. For details, see Jumbo Frames.

    • Features an ultra-high packet forwarding rate (pps).

    • Smaller instance types offer burstable network bandwidth.

    • Network performance scales with the instance type.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

The following table lists the instance types and specifications of the r8ae family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Packet rate (pps)

vTPM support

Connections

Multi-queue

ENIs

Private IPv4/ENI

IPv6 addresses/ENI

Disk IOPS

Disk bandwidth (Gbit/s)

ecs.r8ae.large

2

16

3/15

1,000,000

Yes

Up to 300,000

2

3

6

6

30,000/200,000

2/10

ecs.r8ae.xlarge

4

32

4/15

1,200,000

Yes

Up to 300,000

4

4

15

15

50,000/200,000

2.5/10

ecs.r8ae.2xlarge

8

64

6/15

1,600,000

Yes

Up to 300,000

8

4

15

15

60,000/200,000

3/10

ecs.r8ae.4xlarge

16

128

12/25

3,000,000

Yes

500,000

16

8

30

30

100,000/200,000

6/10

ecs.r8ae.8xlarge

32

256

20/25

6,000,000

Yes

1,000,000

32

8

30

30

200,000/N/A

10/N/A

Note

For ecs.r8ae.large and ecs.r8ae.xlarge instances, you must enable jumbo frames to achieve a burst network bandwidth of 15 Gbit/s. For details, see Jumbo Frames.

r7a memory-optimized instance family

  • Introduction: Powered by the third-generation SHENLONG architecture, this instance family delivers stable and predictable ultra-high performance. It also uses on-chip fast-path acceleration, delivering an order-of-magnitude improvement in storage performance, network performance, and computing stability.

  • Compute:

    • The vCPU-to-memory ratio is 1:8.

    • Processor: AMD EPYC MILAN processors with a base frequency of 2.55 GHz and a single-core turbo frequency of up to 3.5 GHz, which provide stable computing performance.

    • Hyper-Threading is supported and enabled by default. For more information, see Change CPU options.

    • For details on operating system compatibility, see AMD instance and operating system compatibility.

  • Use cases:

    • High-performance databases and in-memory databases.

    • High packet forwarding rate (pps) scenarios, such as live video commenting and telecom data forwarding.

    • Data analytics and mining, and distributed in-memory caches.

    • Hadoop and Spark clusters, and other enterprise applications with high memory requirements.

    • Blockchain applications.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For details on IPv6 communication, see IPv6 communication.

    • Provides an ultra-high packet forwarding rate (pps).

    • Smaller instance types feature burstable network bandwidth.

    • Network performance scales with the instance type.

The following table lists the specifications for the r7a memory-optimized instance family.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Packet rate (pps)

Connections

Multi-queue

ENIs

Private IPv4/ENI

IPv6/ENI

Cloud disk IOPS

Cloud disk bandwidth (Gbit/s)

ecs.r7a.large

2

16

1/Up to 10

900,000

Up to 250,000

2

3

6

6

12,500/Up to 110,000

1/Up to 6

ecs.r7a.xlarge

4

32

1.5/Up to 10

1,000,000

Up to 250,000

4

4

15

15

20,000/Up to 110,000

1.5/Up to 6

ecs.r7a.2xlarge

8

64

2.5/Up to 10

1,600,000

Up to 250,000

8

4

15

15

30,000/Up to 110,000

2/Up to 6

ecs.r7a.4xlarge

16

128

5/Up to 10

2,000,000

300,000

8

8

30

30

60,000/Up to 110,000

3/Up to 6

ecs.r7a.8xlarge

32

256

8/Up to 10

3,000,000

600,000

16

7

30

30

75,000/Up to 110,000

4/Up to 6

ecs.r7a-nps1.8xlarge

32

256

8/Up to 10

3,000,000

600,000

16

7

30

30

75,000/Up to 110,000

4/Up to 6

ecs.r7a.16xlarge

64

512

16/N/A

6,000,000

1,000,000

32

7

30

30

150,000/N/A

8/N/A

ecs.r7a-nps1.16xlarge

64

512

16/N/A

6,000,000

1,000,000

32

8

30

30

150,000/N/A

8/N/A

ecs.r7a.32xlarge

128

1024

32/N/A

12,000,000

2,000,000

32

15

30

30

300,000/N/A

16/N/A

Note

Do not use Ubuntu 16 or Debian 9 images to create instances of this family. The kernels in these operating systems do not support AMD EPYC MILAN processors, and the resulting instances will fail to start.

R7 memory-optimized instance family

  • Introduction: Powered by the third-generation X-Dragon architecture, this instance family delivers stable and predictable ultra-high performance. It also uses on-chip fast-path acceleration to achieve an order-of-magnitude improvement in storage performance, network performance, and compute stability.

  • Use cases:

    • High-performance databases and in-memory databases.

    • High packet-per-second (PPS) workloads, such as video live commenting and telecom service forwarding.

    • Data analytics and mining, and distributed in-memory caches.

    • Hadoop and Spark clusters, and other enterprise applications with high memory requirements.

    • Applications that require secure and trusted computing.

  • Compute:

    • The vCPU-to-memory ratio is 1:8.

    • Processor: Third-generation Intel® Xeon® Scalable processors (Ice Lake) with a base frequency of 2.7 GHz and an all-core turbo frequency of 3.5 GHz.

    • You can enable or disable Hyper-Threading.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Delivers an ultra-high packet forwarding rate (PPS).

    • Smaller instance types feature burstable network bandwidth.

    • Network performance scales with the instance type.

  • Security:

The following table lists the instance types and specifications for the r7 instance family.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

vTPM support

Connections

NIC queues

ENIs

Private IPv4s per ENI

IPv6s per ENI

Max data disks

Cloud disk IOPS (baseline/burst)

Cloud disk bandwidth (Gbit/s)

ecs.r7.large

2

16

2/Up to 12.5

1,100,000

Yes

Up to 500,000

2

3

6

6

8

20,000/Up to 160,000

1.5/Up to 10

ecs.r7.xlarge

4

32

3/Up to 12.5

1,100,000

Yes

Up to 500,000

4

4

15

15

8

40,000/Up to 160,000

2/Up to 10

ecs.r7.2xlarge

8

64

5/Up to 15

1,600,000

Yes

Up to 500,000

8

4

15

15

16

50,000/Up to 160,000

3/Up to 10

ecs.r7.3xlarge

12

96

8/Up to 15

2,400,000

Yes

Up to 500,000

8

8

15

15

24

70,000/Up to 160,000

4/Up to 10

ecs.r7.4xlarge

16

128

10/Up to 25

3,000,000

Yes

500,000

8

8

30

30

24

80,000/Up to 160,000

5/Up to 10

ecs.r7.6xlarge

24

192

12/Up to 25

4,500,000

Yes

550,000

12

8

30

30

24

110,000/160,000

6/10

ecs.r7.8xlarge

32

256

16/Up to 32

6,000,000

Yes

600,000

16

8

30

30

24

160,000/N/A

10/N/A

ecs.r7.16xlarge

64

512

32/N/A

12,000,000

Yes

1,200,000

32

8

30

30

32

360,000/N/A

20/N/A

ecs.r7.32xlarge

128

1024

64/N/A

24,000,000

Yes

2,400,000

32

15

30

30

32

600,000/N/A

32/N/A

R6 memory-optimized instance family

  • Introduction: This instance family is built on the SHENLONG architecture, which offloads virtualization functions to dedicated hardware. This reduces virtualization overhead and delivers stable, predictable, and ultra-high performance.

  • Use cases:

    • High packet-per-second (pps) workloads, such as video bullet comments and telecom data forwarding.

    • High-performance databases and in-memory databases.

    • Data analytics and mining, and distributed in-memory cache.

    • Hadoop and Spark clusters, and other memory-intensive enterprise applications.

  • Compute:

    • The vCPU-to-memory ratio is 1:8.

    • Processor: Intel® Xeon® Platinum 8269CY (Cascade Lake) processors with a base frequency of 2.5 GHz and a turbo frequency of 3.2 GHz, providing stable computing performance.

    • Hyper-Threading is supported and enabled by default. For more information, see Change CPU options.

    Note

    This instance family does not support FreeBSD 13.2 or earlier due to compatibility issues. Use FreeBSD 13.3 or later.

  • Storage:

    • All instances of this family are I/O optimized.

    • Supported cloud disk types: ESSD, ESSD AutoPL disks, Regional ESSD, standard SSD, and ultra disk. For more information, see Elastic Block Storage.

      Note

      The maximum performance of disks varies based on the instance families. A single instance of this instance family can deliver up to 200,000 IOPS.

    • Storage I/O performance scales with the instance type. For more information, see Storage I/O performance.

  • Network:

    • IPv4 and IPv6 are supported. For information about IPv6 communication, see IPv6 communication.

    • Provides an ultra-high packet forwarding rate (pps).

    • Network performance scales with the instance type.

The following table lists the specifications for the r6 instance family.

Instance type

vCPUs

Memory (GiB)

Baseline/burst bandwidth (Gbit/s)

Packet rate (pps)

Connections

NIC queues

ENIs

Private IPv4s per ENI

IPv6s per ENI

Baseline IOPS

Baseline bandwidth (Gbit/s)

ecs.r6.large

2

16

1/Up to 3

300,000

Up to 250,000

2

2

6

1

10,000

1

ecs.r6.xlarge

4

32

1.5/Up to 5

500,000

Up to 250,000

4

3

10

1

20,000

1.5

ecs.r6.2xlarge

8

64

2.5/Up to 8

800,000

Up to 250,000

8

4

10

1

25,000

2

ecs.r6.3xlarge

12

96

4/Up to 10

900,000

Up to 250,000

8

6

10

1

30,000

2.5

ecs.r6.4xlarge

16

128

5/Up to 10

1,000,000

300,000

8

8

20

1

40,000

3

ecs.r6.6xlarge

24

192

7.5/Up to 10

1,500,000

450,000

12

8

20

1

50,000

4

ecs.r6.8xlarge

32

256

10/N/A

2,000,000

600,000

16

8

20

1

60,000

5

ecs.r6.13xlarge

52

384

12.5/N/A

3,000,000

900,000

32

7

20

1

100,000

8

ecs.r6.26xlarge

104

768

25/N/A

6,000,000

1,800,000

32

15

20

1

200,000

16

Memory-optimized instance family r6a

  • Introduction: This instance family uses the SHENLONG architecture to offload virtualization functions to dedicated hardware. This design reduces virtualization overhead and delivers stable, predictable performance.

  • Use cases: Video encoding and decoding, networking scenarios that require a high packet forwarding rate (pps), in-memory databases, memory-intensive enterprise applications such as Hadoop and Spark, and development and testing (for example, DevOps).

  • Compute:

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Provides an ultra-high packet forwarding rate (pps).

    • Network performance scales with the instance type.

The following table lists the specifications for the r6a instance family.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (baseline/burst)

Packet rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses

IPv6 addresses

Disk baseline IOPS

Disk baseline bandwidth

ecs.r6a.large

2

16

1/10

900,000

Up to 250,000

2

2

6

1

12,500

1

ecs.r6a.xlarge

4

32

1.5/10

1,000,000

Up to 250,000

4

3

15

1

20,000

1.5

ecs.r6a.2xlarge

8

64

2.5/10

1,600,000

Up to 250,000

8

4

15

1

30,000

2

ecs.r6a.4xlarge

16

128

5/10

2,000,000

300,000

8

8

30

1

60,000

3.1

ecs.r6a.8xlarge

32

256

8/10

3,000,000

600,000

16

7

30

1

75,000

4.1

ecs.r6a.16xlarge

64

512

16/N/A

6,000,000

1,000,000

32

8

30

1

150,000

8.2

r6e: Enhanced memory-optimized instance family

  • Introduction: This instance family is built on the third-generation SHENLONG architecture, which offloads virtualization features to dedicated hardware. This approach reduces virtualization overhead and delivers stable, predictable, and extremely high performance. Chip-level fast-path acceleration also boosts storage and network performance by an order of magnitude and enhances compute stability.

  • Use cases:

    • High packet-per-second (PPS) workloads, such as video bullet comments and telecom service forwarding.

    • High-performance databases and in-memory databases.

    • Data analytics and mining, and distributed in-memory caches.

    • Hadoop and Spark clusters, and other enterprise applications with high memory requirements.

  • Compute:

    • The processor-to-memory ratio is approximately 1:8.

    • Processor: Intel® Xeon® Platinum 8269CY (Cascade Lake) with a base frequency of 2.5 GHz and a turbo frequency of 3.2 GHz, ensuring stable compute performance.

    • Hyper-Threading is supported and enabled by default. For more information, see Change CPU options.

    Note

    This instance family is incompatible with FreeBSD 13.2 and earlier versions. Use FreeBSD 13.3 or a later version.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Offers an extremely high packet forwarding rate (PPS).

      Note

      Network performance varies based on the instance families. For higher concurrent connection and network packet forwarding capabilities, we recommend that you use the g7ne instance family.

The following table lists the instance types and specifications of the r6e instance family.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Packet rate (pps)

Connections

NIC queues

ENIs

Private IPv4s/ENI

IPv6s/ENI

Disk baseline IOPS

Baseline bandwidth (Gbit/s)

ecs.r6e.large

2

16

1.2/Up to 10

900,000

Up to 250,000

2

3

6

1

20,000

1

ecs.r6e.xlarge

4

32

2/Up to 10

1,000,000

Up to 250,000

4

4

15

1

40,000

1.5

ecs.r6e.2xlarge

8

64

3/Up to 10

1,600,000

Up to 250,000

8

4

15

1

50,000

2

ecs.r6e.4xlarge

16

128

6/Up to 10

3,000,000

300,000

8

8

30

1

80,000

3

ecs.r6e.8xlarge

32

256

10/N/A

6,000,000

600,000

16

8

30

1

150,000

5

ecs.r6e.13xlarge

52

384

16/N/A

9,000,000

1,000,000

32

7

30

1

240,000

8

ecs.r6e.26xlarge

104

768

32/N/A

24,000,000

1,800,000

32

15

30

1

480,000

16

R5 memory-optimized instance family

  • Scenarios:

    • Workloads that require a high packet forwarding rate (pps), such as live video commenting and telecom service forwarding.

    • High-performance databases and in-memory databases.

    • Data analytics and mining, and distributed in-memory cache.

    • Hadoop and Spark clusters, and other memory-intensive enterprise applications.

  • Compute:

    • vCPU-to-memory ratio of 1:8.

    • Processor: 2.5 GHz Intel® Xeon® Platinum 8163 (Skylake) or Intel® Xeon® Platinum 8269CY (Cascade Lake) processors provide consistent computing performance.

      Note

      Instances in this instance family may be deployed on different server platforms. If your workloads require all instances to be deployed on the same server platform, use r6, r6e, or r7 instances.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For more information about IPv6, see IPv6 communication.

    • Ultra-high packet forwarding rate (pps).

    • The network performance of an instance scales with its instance type.

r5 includes the following instance types and specifications:

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

PPS

NIC queues

ENIs

IPv4 addresses per ENI

IPv6 addresses per ENI

ecs.r5.large

2

16

1

300,000

2

2

6

1

ecs.r5.xlarge

4

32

1.5

500,000

2

3

10

1

ecs.r5.2xlarge

8

64

2.5

800,000

4

4

10

1

ecs.r5.3xlarge

12

96

4

900,000

4

6

10

1

ecs.r5.4xlarge

16

128

5

1,000,000

4

8

20

1

ecs.r5.6xlarge

24

192

7.5

1,500,000

6

8

20

1

ecs.r5.8xlarge

32

256

10

2,000,000

8

8

20

1

ecs.r5.16xlarge

64

512

20

4,000,000

16

8

20

1

se1ne: Memory-optimized, network-enhanced instance family

  • Use cases:

    • High-performance workloads with a high packet forwarding rate (pps), such as video livestreaming with on-screen comments and telecom service forwarding

    • High-performance databases and in-memory databases

    • Data analytics and mining, and distributed memory caches

    • Hadoop and Spark clusters, and other enterprise applications with high memory requirements

  • Compute:

    • The vCPU-to-memory ratio is 1:8.

    • Processor: 2.5 GHz Intel® Xeon® E5-2682 v4 (Broadwell), Platinum 8163 (Skylake), or 8269CY (Cascade Lake) processors for stable computing performance.

      Note

      se1ne instances may be deployed on different server platforms. If your workloads require all instances to run on the same server platform, use r6, r6e, or r7.

  • Storage:

    • All se1ne instances are I/O optimized instances.

    • Supported cloud disk types: standard SSDs and ultra disks. For more information, see block storage overview.

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Offers an ultra-high packet forwarding rate (pps).

    • Network performance scales with the instance type.

The following table lists the se1ne instance types and their specifications.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Packet rate (pps)

NIC queues

ENIs

Private IPv4 addresses

IPv6 addresses

ecs.se1ne.large

2

16

1

300,000

2

2

6

1

ecs.se1ne.xlarge

4

32

1.5

500,000

2

3

10

1

ecs.se1ne.2xlarge

8

64

2

1,000,000

4

4

10

1

ecs.se1ne.3xlarge

12

96

2.5

1,300,000

4

6

10

1

ecs.se1ne.4xlarge

16

128

3

1,600,000

4

8

20

1

ecs.se1ne.6xlarge

24

192

4.5

2,000,000

6

8

20

1

ecs.se1ne.8xlarge

32

256

6

2,500,000

8

8

20

1

ecs.se1ne.14xlarge

56

480

10

4,500,000

14

8

20

1

Se1 memory-optimized instance family

  • Use cases:

    • High-performance databases and in-memory databases.

    • Data analytics, data mining, and distributed in-memory caches.

    • Hadoop, Spark, and other memory-intensive enterprise applications.

  • Compute:

    • The vCPU-to-memory ratio is 1:8.

    • Processor: 2.5 GHz Intel® Xeon® E5-2682 v4 (Broadwell), Intel® Xeon® Platinum 8163 (Skylake), or Intel® Xeon® 8269CY (Cascade Lake) processors.

  • Storage:

    • I/O-optimized instances.

    • Supported cloud disk types: Standard SSD and Ultra Disk. For more information, see Block Storage.

  • Network:

    • Supports only IPv4.

    • Network performance scales with the instance type.

The following table lists the specifications of the se1 instance family.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Packet forwarding rate (pps)

Multi-queue

ENIs

Private IPv4 addresses per ENI

ecs.se1.large

2

16

0.5

100,000

1

2

6

ecs.se1.xlarge

4

32

0.8

200,000

1

3

10

ecs.se1.2xlarge

8

64

1.5

400,000

1

4

10

ecs.se1.4xlarge

16

128

3

500,000

2

8

20

ecs.se1.8xlarge

32

256

6

800,000

3

8

20

ecs.se1.14xlarge

56

480

10

1,200,000

4

8

20

u2a, universal instance family

  • Introduction: u2a instances use Alibaba Cloud's new Cloud Infrastructure Processing Unit (CIPU) architecture. They are compatible with multiple generations of AMD EPYC processors, including AMD Turin processors, and provides enterprise-grade computing power.

  • Use cases:

    • Small and medium-sized databases, such as Redis and MySQL

    • Application servers

    • Middleware, such as MQ and Kafka

    • Websites or network access layers, such as Apache and Nginx

    • Other internal enterprise systems, such as development and testing environments and email systems

  • Compute:

    • Supports instance types with vCPU-to-memory ratios of 1:1, 1:2, and 1:4.

    • Processor: AMD EPYC processor with a turbo frequency of up to 3.7 GHz.

    • For more information about operating system compatibility, see AMD instance and operating system compatibility.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Supports Elastic RDMA Interface (ERI). For more information about how to use ERI, see Enable eRDMA on enterprise-level instances.

    • Supports Jumbo Frames.

    • The network performance of an instance depends on its instance type. A larger instance type provides better network performance.

u2a instance types

Instance type

vCPUs

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.u2a-c1m1.large

2

2

1.2/Up to 12.5

Up to 900,000

Up to 250,000

2

2

6

6

Up to 110,000

1.2/Up to 10

ecs.u2a-c1m2.large

2

4

1.2/Up to 12.5

Up to 900,000

Up to 250,000

2

2

6

6

Up to 110,000

1.2/Up to 10

ecs.u2a-c1m4.large

2

8

1.2/Up to 12.5

Up to 900,000

Up to 250,000

2

2

6

6

Up to 110,000

1.2/Up to 10

ecs.u2a-c1m1.xlarge

4

4

2/Up to 12.5

Up to 1,000,000

Up to 250,000

4

2

15

15

Up to 110,000

1.6/Up to 10

ecs.u2a-c1m2.xlarge

4

8

2/Up to 12.5

Up to 1,000,000

Up to 250,000

4

2

15

15

Up to 110,000

1.6/Up to 10

ecs.u2a-c1m4.xlarge

4

16

2/Up to 12.5

Up to 1,000,000

Up to 250,000

4

2

15

15

Up to 110,000

1.6/Up to 10

ecs.u2a-c1m2.2xlarge

8

16

3.2/Up to 12.5

Up to 1,600,000

Up to 250,000

8

2

15

15

Up to 110,000

2/Up to 10

ecs.u2a-c1m4.2xlarge

8

32

3.2/Up to 12.5

Up to 1,600,000

Up to 250,000

8

2

15

15

Up to 110,000

2/Up to 10

ecs.u2a-c1m2.4xlarge

16

32

6/Up to 12.5

Up to 2,000,000

Up to 250,000

16

4

30

30

Up to 110,000

3/Up to 10

ecs.u2a-c1m4.4xlarge

16

64

6/Up to 12.5

Up to 2,000,000

Up to 250,000

16

4

30

30

Up to 110,000

3/Up to 10

ecs.u2a-c1m2.6xlarge

24

48

7/Up to 12.5

Up to 2,500,000

Up to 360,000

16

8

30

30

Up to 110,000

3.5/Up to 10

ecs.u2a-c1m4.6xlarge

24

96

7/Up to 12.5

Up to 2,500,000

Up to 360,000

16

8

30

30

Up to 110,000

3.5/Up to 10

ecs.u2a-c1m2.8xlarge

32

64

8/Up to 25

Up to 3,000,000

500,000

16

8

30

30

Up to 110,000

4/Up to 10

ecs.u2a-c1m4.8xlarge

32

128

8/Up to 25

Up to 3,000,000

500,000

16

8

30

30

Up to 110,000

4/Up to 10

ecs.u2a-c1m2.12xlarge

48

96

13/Up to 25

4,000,000

600,000

24

8

30

30

Up to 110,000

6.5/Up to 10

ecs.u2a-c1m4.12xlarge

48

192

13/Up to 25

4,000,000

600,000

24

8

30

30

Up to 110,000

6.5/Up to 10

ecs.u2a-c1m2.16xlarge

64

128

16/Up to 25

5,000,000

800,000

32

8

30

30

Up to 110,000

8/Up to 10

ecs.u2a-c1m4.16xlarge

64

256

16/Up to 25

5,000,000

800,000

32

8

30

30

Up to 110,000

8/Up to 10

u2i, universal instance family

  • Introduction: u2i instances use Alibaba Cloud's new CIPU architecture. They are compatible with servers from multiple generations and support the latest 5th and 6th generation Intel Xeon platforms.

  • Use cases:

    • Small and medium-sized enterprise applications

    • Websites and application servers

    • Data analytics and computing

    • Small and medium-sized database systems, caches, and search clusters

  • Compute:

    • CPU-to-memory ratios: 1:1, 1:2, 1:4, and 1:8.

    • Processor: Intel ® Xeon ® Platinum Scalable processor.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • The network performance of an instance depends on its computing specifications. A higher specification provides better network performance.

u2i instance types

Instance type

vCPUs

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.u2i-c1m1.large

2

2

2/Up to 15

900,000

Up to 300,000

2

3

6

6

20,000/Up to 200,000

1.5/10

ecs.u2i-c1m1.xlarge

4

4

3/Up to 15

1,100,000

Up to 300,000

2

4

15

15

40,000/Up to 200,000

2/10

ecs.u2i-c1m1.2xlarge

8

8

5/Up to 15

1,300,000

Up to 300,000

4

4

15

15

50,000/Up to 200,000

3/10

ecs.u2i-c1m1.3xlarge

12

12

8/Up to 15

1,800,000

Up to 300,000

6

8

15

15

60,000/Up to 200,000

4/10

ecs.u2i-c1m1.4xlarge

16

16

10/Up to 25

2,400,000

Up to 350,000

8

8

30

30

80,000/Up to 200,000

5/10

ecs.u2i-c1m1.6xlarge

24

24

12/Up to 25

3,000,000

Up to 400,000

12

8

30

30

100,000/Up to 200,000

6/10

ecs.u2i-c1m1.8xlarge

32

32

16/Up to 32

4,000,000

Up to 600,000

16

8

30

30

160,000/Up to 300,000

8/12

ecs.u2i-c1m2.large

2

4

2/Up to 15

900,000

Up to 300,000

2

3

6

6

20,000/Up to 200,000

1.5/10

ecs.u2i-c1m2.xlarge

4

8

3/Up to 15

1,100,000

Up to 300,000

2

4

15

15

40,000/Up to 200,000

2/10

ecs.u2i-c1m2.2xlarge

8

16

5/Up to 15

1,300,000

Up to 300,000

4

4

15

15

50,000/Up to 200,000

3/10

ecs.u2i-c1m2.3xlarge

12

24

8/Up to 15

1,800,000

Up to 300,000

6

8

15

15

60,000/Up to 200,000

4/10

ecs.u2i-c1m2.4xlarge

16

32

10/Up to 25

2,400,000

Up to 350,000

8

8

30

30

80,000/Up to 200,000

5/10

ecs.u2i-c1m2.6xlarge

24

48

12/Up to 25

3,000,000

Up to 400,000

12

8

30

30

100,000/Up to 200,000

6/10

ecs.u2i-c1m2.8xlarge

32

64

16/Up to 32

4,000,000

Up to 600,000

16

8

30

30

160,000/Up to 300,000

8/12

ecs.u2i-c1m4.large

2

8

2/Up to 15

900,000

Up to 300,000

2

3

6

6

20,000/Up to 200,000

1.5/10

ecs.u2i-c1m4.xlarge

4

16

3/Up to 15

1,100,000

Up to 300,000

2

4

15

15

40,000/Up to 200,000

2/10

ecs.u2i-c1m4.2xlarge

8

32

5/Up to 15

1,300,000

Up to 300,000

4

4

15

15

50,000/Up to 200,000

3/10

ecs.u2i-c1m4.3xlarge

12

48

8/Up to 15

1,800,000

Up to 300,000

6

8

15

15

60,000/Up to 200,000

4/10

ecs.u2i-c1m4.4xlarge

16

64

10/Up to 25

2,400,000

Up to 350,000

8

8

30

30

80,000/Up to 200,000

5/10

ecs.u2i-c1m4.6xlarge

24

96

12/Up to 25

3,000,000

Up to 400,000

12

8

30

30

100,000/Up to 200,000

6/10

ecs.u2i-c1m4.8xlarge

32

128

16/Up to 32

4,000,000

Up to 600,000

16

8

30

30

160,000/Up to 300,000

8/12

ecs.u2i-c1m8.large

2

16

2/Up to 15

900,000

Up to 300,000

2

3

6

6

20,000/Up to 200,000

1.5/10

ecs.u2i-c1m8.xlarge

4

32

3/Up to 15

1,100,000

Up to 300,000

2

4

15

15

40,000/Up to 200,000

2/10

ecs.u2i-c1m8.2xlarge

8

64

5/Up to 15

1,300,000

Up to 300,000

4

4

15

15

50,000/Up to 200,000

3/10

ecs.u2i-c1m8.3xlarge

12

96

8/Up to 15

1,800,000

Up to 300,000

6

8

15

15

60,000/Up to 200,000

4/10

ecs.u2i-c1m8.4xlarge

16

128

10/Up to 25

2,400,000

Up to 350,000

8

8

30

30

80,000/Up to 200,000

5/10

ecs.u2i-c1m8.6xlarge

24

192

12/Up to 25

3,000,000

Up to 400,000

12

8

30

30

100,000/Up to 200,000

6/10

ecs.u2i-c1m8.8xlarge

32

256

16/Up to 32

4,000,000

Up to 600,000

16

8

30

30

160,000/Up to 300,000

8/12

u1, universal instance family

  • Use cases:

    • Small and medium-sized enterprise applications

    • Websites and application servers

    • Data analytics and computing

    • Small and medium-sized database systems, caches, and search clusters

  • Compute:

    • CPU-to-memory ratios: 1:1, 1:2, 1:4, and 1:8.

    • Processor: Intel ® Xeon ® Platinum Scalable processor

    Note

    Instances of this family are randomly deployed on different server platforms when they are created and may be migrated to different server platforms during their lifecycle. u1 instances use technical means to improve compatibility between different platforms. However, performance may vary significantly between platforms. If you require consistent performance, we recommend that you use g9i, c9i, or r9i instances.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • The network performance of an instance depends on its computing specifications. A higher specification provides better network performance.

u1 instance types

Instance type

vCPUs

Memory (GiB)

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.u1-c1m1.large

2

2

1

300,000

Up to 250,000

2

2

6

2

10,000

1

ecs.u1-c1m2.large

2

4

1

300,000

Up to 250,000

2

2

6

2

10,000

1

ecs.u1-c1m4.large

2

8

1

300,000

Up to 250,000

2

2

6

2

10,000

1

ecs.u1-c1m8.large

2

16

1

300,000

Up to 250,000

2

2

6

2

10,000

1

ecs.u1-c1m1.xlarge

4

4

1.5

500,000

Up to 250,000

2

3

10

2

20,000

1.5

ecs.u1-c1m2.xlarge

4

8

1.5

500,000

Up to 250,000

2

3

10

2

20,000

1.5

ecs.u1-c1m4.xlarge

4

16

1.5

500,000

Up to 250,000

2

3

10

2

20,000

1.5

ecs.u1-c1m8.xlarge

4

32

1.5

500,000

Up to 250,000

2

3

10

2

20,000

1.5

ecs.u1-c1m1.2xlarge

8

8

2.5

800,000

Up to 250,000

4

4

10

2

25,000

2

ecs.u1-c1m2.2xlarge

8

16

2.5

800,000

Up to 250,000

4

4

10

2

25,000

2

ecs.u1-c1m4.2xlarge

8

32

2.5

800,000

Up to 250,000

4

4

10

2

25,000

2

ecs.u1-c1m8.2xlarge

8

64

2.5

800,000

Up to 250,000

4

4

10

2

25,000

2

ecs.u1-c1m1.3xlarge

12

12

4

900,000

Up to 250,000

4

6

10

2

30,000

2.5

ecs.u1-c1m2.3xlarge

12

24

4

900,000

Up to 250,000

4

6

10

2

30,000

2.5

ecs.u1-c1m4.3xlarge

12

48

4

900,000

Up to 250,000

4

6

10

2

30,000

2.5

ecs.u1-c1m8.3xlarge

12

96

4

900,000

Up to 250,000

4

6

10

2

30,000

2.5

ecs.u1-c1m1.4xlarge

16

16

5

1,000,000

Up to 300,000

4

8

20

2

40,000

3

ecs.u1-c1m2.4xlarge

16

32

5

1,000,000

Up to 300,000

4

8

20

2

40,000

3

ecs.u1-c1m4.4xlarge

16

64

5

1,000,000

Up to 300,000

4

8

20

2

40,000

3

ecs.u1-c1m8.4xlarge

16

128

5

1,000,000

Up to 300,000

4

8

20

2

40,000

3

ecs.u1-c1m1.8xlarge

32

32

10

2,000,000

Up to 300,000

8

8

20

2

60,000

5

ecs.u1-c1m2.8xlarge

32

64

10

2,000,000

Up to 300,000

8

8

20

2

60,000

5

ecs.u1-c1m4.8xlarge

32

128

10

2,000,000

Up to 300,000

8

8

20

2

60,000

5

ecs.u1-c1m8.8xlarge

32

256

10

2,000,000

Up to 300,000

8

8

20

2

60,000

5

Note
  • Exceptions may occur when you deploy Data Plane Development Kit (DPDK) applications on u1 instances. To resolve this issue, use Virtual Function I/O (VFIO) drivers to replace Userspace I/O (UIO) drivers. For more information, see Use VFIO drivers to replace UIO drivers.

  • For frequently asked questions about Universal instances, see U1 instance FAQ.

d3s, storage-intensive big data instance family

Features:

  • This instance family is equipped with 12-TB, large-capacity, high-throughput local SATA HDDs and can provide a maximum network bandwidth of 64 Gbit/s between instances.

  • Supported scenarios:

    • Big data computing and storage business scenarios in which services such as Hadoop MapReduce, HDFS, Hive, and HBase are used

    • Machine learning scenarios such as Spark in-memory computing and MLlib

    • Search and log data processing scenarios in which solutions such as Elasticsearch and Kafka are used

  • This instance family supports online replacement and hot swapping of damaged disks to prevent instance shutdown.

    If a local disk fails, you receive a system event. You can handle the system event by initiating the process of repairing the damaged disk. For more information, see Maintenance and system events for local disk instances.

    Important

    After you initiate the process of repairing the damaged disk, data stored on the damaged disk cannot be restored.

  • Compute:

    • Uses 2.7 GHz Intel® Xeon® Scalable (Ice Lake) processors that deliver an all-core turbo frequency of 3.5 GHz to provide consistent computing performance.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports only ESSDs and ESSD AutoPL disks.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

d3s includes the instance types and metric data listed in the following table.

Instance type

vCPUs

Memory (GiB)

Local storage

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Disk baseline/burst bandwidth (Gbit/s)

ecs.d3s.2xlarge

8

32

4 * 11,918 GB (4 * 11,100 GiB)

10/burstable up to 15

2,000,000

3/burstable up to 5

ecs.d3s.4xlarge

16

64

8 * 11,918 GB (8 * 11,100 GiB)

25/none

3,000,000

5/none

ecs.d3s.8xlarge

32

128

16 * 11,918 GB (16 * 11,100 GiB)

40/none

6,000,000

8/none

ecs.d3s.12xlarge

48

192

24 * 11,918 GB (24 * 11,100 GiB)

60/none

9,000,000

12/none

ecs.d3s.16xlarge

64

256

32 * 11,918 GB (32 * 11,100 GiB)

80/none

12,000,000

16/none

d3c, compute-intensive big data instance family

Features:

  • This instance family is equipped with high-capacity and high-throughput local disks and can provide a maximum bandwidth of 40 Gbit/s between instances.

  • Supported scenarios:

    • Big data computing and storage business scenarios in which services such as Hadoop MapReduce, HDFS, Hive, and HBase are used

    • Scenarios in which EMR JindoFS and Object Storage Service (OSS) are used in combination to separately store hot and cold data and decouple storage from computing

    • Machine learning scenarios such as Spark in-memory computing and MLlib

    • Search and log data processing scenarios in which solutions such as Elasticsearch and Kafka are used

  • This instance family supports online replacement and hot swapping of damaged disks to prevent instance shutdown.

    If a local disk fails, you receive a system event. You can handle the system event by initiating the process of repairing the damaged disk. For more information, see Maintenance and system events for local disk instances.

    Important

    After you initiate the process of repairing the damaged disk, data stored on the damaged disk cannot be restored.

  • Compute:

    • Uses third-generation 2.9 GHz Intel® Xeon® Scalable (Ice Lake) processors that deliver an all-core turbo frequency of 3.5 GHz to provide consistent computing performance.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports only ESSDs and ESSD AutoPL disks.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

d3c includes the instance types and metric data listed in the following table.

Instance type

vCPUs

Memory (GiB)

Local storage

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.d3c.3xlarge

14

56.0

1 * 13,743 GB (1 * 12,800 GiB)

8/burstable up to 10

1,600,000

40,000/none

3/none

ecs.d3c.7xlarge

28

112.0

2 * 13,743 GB (2 * 12,800 GiB)

16/burstable up to 25

2,500,000

50,000/none

4/none

ecs.d3c.14xlarge

56

224.0

4 * 13,743 GB (4 * 12,800 GiB)

40/none

5,000,000

100,000/none

8/none

Note

This instance family supports only Linux images. When you create an instance of this instance family, select a Linux image.

d2c, compute-intensive big data instance family

Features:

  • This instance family is equipped with high-capacity and high-throughput local SATA HDDs and can provide a maximum bandwidth of 35 Gbit/s between instances.

  • Supported scenarios:

    • Big data computing and storage business scenarios in which services such as Hadoop MapReduce, HDFS, Hive, and HBase are used

    • Scenarios in which EMR JindoFS and OSS are used in combination to separately store hot and cold data and decouple storage from computing

    • Machine learning scenarios such as Spark in-memory computing and MLlib

    • Search and log data processing scenarios in which solutions such as Elasticsearch and Kafka are used

  • This instance family supports online replacement and hot swapping of damaged disks to prevent instance shutdown.

    If a local disk fails, you receive a system event. You can handle the system event by initiating the process of repairing the damaged disk. For more information, see Maintenance and system events for local disk instances.

    Important

    After you initiate the process of repairing the damaged disk, data stored on the damaged disk cannot be restored.

  • Compute:

    • Uses 2.5 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processors.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports enhanced SSDs (ESSDs), ESSD AutoPL disks, standard SSDs, and ultra disks.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

d2c includes the instance types and metric data listed in the following table.

Instance type

vCPUs

Memory (GiB)

Local storage

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (pps)

ecs.d2c.6xlarge

24

88.0

3 * 3,972 GB (3 * 3,700 GiB)

12.0

1,600,000

ecs.d2c.12xlarge

48

176.0

6 * 3,972 GB (6 * 3,700 GiB)

20.0

2,000,000

ecs.d2c.24xlarge

96

352.0

12 * 3,972 GB (12 * 3,700 GiB)

35.0

4,500,000

d2s, storage-intensive big data instance family

Features:

  • This instance family is equipped with high-capacity and high-throughput local SATA HDDs and can provide a maximum bandwidth of 35 Gbit/s between instances.

  • Supported scenarios:

    • Big data computing and storage business scenarios in which services such as Hadoop MapReduce, HDFS, Hive, and HBase are used

    • Machine learning scenarios such as Spark in-memory computing and MLlib

    • Search and log data processing scenarios in which solutions such as Elasticsearch and Kafka are used

  • This instance family supports online replacement and hot swapping of damaged disks to prevent instance shutdown.

    If a local disk fails, you receive a system event. You can handle the system event by initiating the process of repairing the damaged disk. For more information, see Maintenance and system events for local disk instances.

    Important

    After you initiate the process of repairing the damaged disk, data stored on the damaged disk cannot be restored.

  • Compute:

    • Uses 2.5 GHz Intel® Xeon® Platinum 8163 (Skylake) processors.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports ESSDs, ESSD AutoPL disks, standard SSDs, and ultra disks.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

d2s includes the instance types and metric data listed in the following table.

Instance type

vCPUs

Memory (GiB)

Local storage

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (pps)

ecs.d2s.5xlarge

20

88.0

8 * 7,838 GB (8 * 7,300 GiB)

12.0

1,600,000

ecs.d2s.10xlarge

40

176.0

15 * 7,838 GB (15 * 7,300 GiB)

20.0

2,000,000

ecs.d2s.20xlarge

80

352.0

30 * 7,838 GB (30 * 7,300 GiB)

35.0

4,500,000

d1ne, network-enhanced big data instance family

Features:

  • This instance family is equipped with high-capacity and high-throughput local SATA HDDs and can provide a maximum bandwidth of 35 Gbit/s between instances.

  • Supported scenarios:

    • Scenarios in which services such as Hadoop MapReduce, HDFS, Hive, and HBase are used

    • Machine learning scenarios such as Spark in-memory computing and MLlib

    • Search and log data processing scenarios in which solutions such as Elasticsearch are used

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4, which is designed for big data scenarios.

    • Uses 2.5 GHz Intel® Xeon® E5-2682 v4 (Broadwell) or Intel® Xeon® Platinum 8163 (Skylake) processors, providing stable computing performance.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports only standard SSDs and ultra disks.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

d1ne includes the instance types and metric data listed in the following table.

Instance type

vCPUs

Memory (GiB)

Local storage

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (pps)

ecs.d1ne.2xlarge

8

32.0

4 * 5,905 GB (4 * 5,500 GiB)

6.0

1,000,000

ecs.d1ne.4xlarge

16

64.0

8 * 5,905 GB (8 * 5,500 GiB)

12.0

1,600,000

ecs.d1ne.6xlarge

24

96.0

12 * 5,905 GB (12 * 5,500 GiB)

16.0

2,000,000

ecs.d1ne-c8d3.8xlarge

32

128.0

12 * 5,905 GB (12 * 5,500 GiB)

20.0

2,000,000

ecs.d1ne.8xlarge

32

128.0

16 * 5,905 GB (16 * 5,500 GiB)

20.0

2,500,000

ecs.d1ne-c14d3.14xlarge

56

160.0

12 * 5,905 GB (12 * 5,500 GiB)

35.0

4,500,000

ecs.d1ne.14xlarge

56

224.0

28 * 5,905 GB (28 * 5,500 GiB)

35.0

4,500,000

i5g, instance family with local SSDs

  • Introduction: This instance family provides high-performance NVMe SSD local disks with high input/output operations per second (IOPS), high throughput, and low access latency. It leverages the new Alibaba Cloud Cloud Infrastructure Processing Unit (CIPU) architecture and Intel® Xeon® 6 processors to deliver stable computing power and a powerful I/O engine.

  • Use cases: Disk-based key-value (KV) databases such as RocksDB and ClickHouse. E-MapReduce for big data, including hot/cold data layering, storage-compute separation, and data lakes. Search engines such as Elasticsearch.

  • Compute:

    • vCPU-to-memory ratio of 1:4.

    • Processor: Powered by Intel® Xeon® Granite Rapids with a 3.2 GHz base clock speed and a 3.6 GHz all-core turbo frequency for stable computing performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

i5g instance types

Instance type

vCPU

Memory (GiB)

Local storage

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.i5g.8xlarge

32

128

1 × 1,919 GB

(1 × 1,788 GiB)

16/32

10,000,000

200,000/300,000

10/12

ecs.i5g.16xlarge

64

256

1 × 3,839 GB

(1 × 3,576 GiB)

32/None

20,000,000

300,000/None

16/None

i5ge, instance family with local SSDs

  • Introduction: The i5ge instance family is equipped with high-performance NVMe SSD local disks, delivering high IOPS, high throughput, and low access latency. Built on Alibaba Cloud's new CIPU architecture and powered by Intel® Xeon® 6 processors, these instances provide stable computing power and a powerful I/O engine.

  • Use cases: Disk-based key-value (KV) databases such as RocksDB and ClickHouse, big data computing with local cache, and online transactions.

  • Compute:

    • The vCPU-to-memory ratio is 1:6.

    • Processor: Intel® Xeon® Granite Rapids processors with a 3.2 GHz base clock speed and a 3.6 GHz all-core turbo frequency provide stable computing performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

i5ge instance types

Instance type

vCPU

Memory (GiB)

Local storage

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding (PPS)

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.i5ge.3xlarge

12

72

1 × 1,919 GB

(1 × 1,788 GiB)

25/40

4,000,000

80,000/200,000

5/10

ecs.i5ge.6xlarge

24

144

1 × 3,839 GB

(1 × 3,576 GiB)

50/70

8,000,000

120,000/200,000

7.5/10

ecs.i5ge.12xlarge

48

288

2 × 3,839 GB

(2 × 3,576 GiB)

84/None

15,000,000

240,000/None

12/None

ecs.i5ge.24xlarge

96

576

4 × 3,839 GB

(4 × 3,576 GiB)

172/None

30,000,000

300,000/None

20/None

i5e, instance family with local SSDs

  • Introduction: Equipped with high-performance NVMe SSD local disks, this instance family provides high IOPS, high throughput, and low access latency. This family uses the new Alibaba Cloud CIPU architecture and is powered by Intel® Xeon® 6 processors to deliver stable computing power and a powerful I/O engine.

  • Use cases: Relational databases such as MySQL, remote cache services, and cache layer acceleration.

  • Computing:

    • The vCPU-to-memory ratio is 1:8.

    • Processor: Intel® Xeon® Granite Rapids with a 2.9 GHz clock speed and a 3.6 GHz all-core turbo frequency. This processor provides stable computing performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

i5e instance types

Instance type

vCPU

Memory (GiB)

Local storage

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.i5e.2xlarge

8

64

1 × 3840 GB

(1 × 3576 GiB)

20/40

4,000,000

60,000/200,000

4/10

ecs.i5e.4xlarge

16

128

1 × 7680 GB

(1 × 7152 GiB)

40/80

7,000,000

100,000/200,000

6/10

ecs.i5e.8xlarge

32

256

2 × 7680 GB

(2 × 7152 GiB)

80/120

14,000,000

150,000/200,000

10/12

ecs.i5e.16xlarge

64

512

4 × 7680 GB

(4 × 7152 GiB)

160

25,000,000

300,000

16

ecs.i5e.32xlarge

128

1024

8 × 7680 GB

(8 × 7152 GiB)

320

50,000,000

600,000

32

i5, instance family with local SSDs

  • Introduction: The i5 instance family features high-performance NVMe SSD local disks that deliver high IOPS, high throughput, and low access latency. The instances use the new Alibaba Cloud CIPU architecture and are powered by Intel® Xeon® 6 processors. This combination provides stable computing power and a powerful I/O engine.

  • Use cases: Disk-based key-value (KV) databases such as RocksDB and ClickHouse. E-MapReduce for big data with hot and cold data layers. Architectures that separate storage and compute. Data lakes. Other I/O-intensive applications with frequent disk writes, such as message middleware and containers.

  • Compute:

    • Processor-to-memory ratio of 1:8.

    • Processor: Intel® Xeon® Granite Rapids with a base clock speed of 3.4 GHz and an all-core turbo frequency of 3.8 GHz. This ensures stable computing performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

i5 instance types

Instance type

vCPU

Memory (GiB)

Local storage

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding (PPS)

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.i5.xlarge

4

32

1 × 960 GB

(1 × 894 GiB)

10/20

2,000,000

40,000/200,000

2/10

ecs.i5.2xlarge

8

64

1 × 1,919 GB

(1 × 1,788 GiB)

20/40

4,000,000

60,000/200,000

4/10

ecs.i5.4xlarge

16

128

1 × 3,839 GB

(1 × 3,576 GiB)

40/80

7,000,000

100,000/200,000

6/10

ecs.i5.8xlarge

32

256

2 × 3,839 GB

(2 × 3,576 GiB)

80/120

14,000,000

150,000/200,000

10/12

ecs.i5.12xlarge

48

384

3 × 3,839 GB

(3 × 3,576 GiB)

120

20,000,000

240,000

12

ecs.i5.16xlarge

64

512

4 × 3,839 GB

(4 × 3,576 GiB)

160

27,000,00

300,000

16

i4, instance family with local SSDs

  • Introduction: This instance family is equipped with high-performance local NVMe SSDs that deliver high IOPS, high I/O throughput, and low latency.

  • Use cases: OLTP and high-performance relational databases, NoSQL databases such as Cassandra and MongoDB, and search scenarios that use solutions such as Elasticsearch.

  • Compute:

    • Uses 2.7 GHz Intel® Xeon® Scalable (Ice Lake) processors that deliver an all-core turbo frequency of 3.5 GHz to provide consistent computing performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Is compatible with specific operating systems. For more information, see Compatibility between the i4 instance types and operating systems.

i4 instance types

Instance type

vCPUs

Memory (GiB)

Local storage

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.i4.large

2

16

1 * 479 GB (1 * 447 GiB)

2.5/15

900,000

20,000/up to 110,000

1.5/6

ecs.i4.xlarge

4

32

1 * 959 GB (1 * 894 GiB)

4/15

1,000,000

40,000/up to 110,000

2/6

ecs.i4.2xlarge

8

64

1 * 1,919 GB (1 * 1,788 GiB)

6/15

1,600,000

50,000/up to 110,000

3/6

ecs.i4.4xlarge

16

128

1 * 3,839 GB (1 * 3,576 GiB)

10/25

3,000,000

80,000/up to 110,000

5/6

ecs.i4.8xlarge

32

256

2 * 3,839 GB (2 * 3,576 GiB)

25/none

6,000,000

150,000/none

8/none

ecs.i4.16xlarge

64

512

4 * 3,839 GB (4 * 3,576 GiB)

50/none

12,000,000

300,000/none

16/none

ecs.i4.32xlarge

128

1,024

8 * 3,839 GB (8 * 3,576 GiB)

100/none

24,000,000

600,000/none

32/none

i4g, instance family with local SSDs

  • Introduction: This instance family is equipped with high-performance local NVMe SSDs that deliver high IOPS, high I/O throughput, and low latency.

  • Use cases: OLTP and high-performance relational databases, E-MapReduce big data scenarios such as tiering of hot and cold data, storage and computing separation, and data lakes, and search scenarios that use solutions such as Elasticsearch.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4, which is designed for high-performance databases.

    • Uses 2.7 GHz Intel® Xeon® Scalable (Ice Lake) processors that deliver an all-core turbo frequency of 3.5 GHz to provide consistent computing performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

i4g instance types

Instance type

vCPUs

Memory (GiB)

Local storage

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.i4g.4xlarge

16

64

1 * 959 GB (1 * 894 GiB)

8/25

3,000,000

100,000

6

ecs.i4g.8xlarge

32

128

1 * 1,919 GB (1 * 1,788 GiB)

16/25

6,000,000

150,000

8

ecs.i4g.16xlarge

64

256

2 * 1,919 GB (2 * 1,788 GiB)

32/none

12,000,000

300,000

16

ecs.i4g.32xlarge

128

512

4 * 1,919 GB (4 * 1,788 GiB)

64/none

24,000,000

600,000

32

Note

This instance family supports only Linux images. When you create an instance of this instance family, you must select a Linux image.

i4r, instance family with local SSDs

  • Introduction: This instance family is equipped with high-performance local NVMe SSDs that deliver high IOPS, high I/O throughput, and low latency.

  • Use cases: OLTP and high-performance relational databases, NoSQL databases such as Cassandra and MongoDB, and search scenarios that use solutions such as Elasticsearch.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:8, which is designed for high-performance databases. This instance family is the most cost-effective instance family that is suitable for scenarios such as hot data tiering and data lakes.

    • Uses 2.7 GHz Intel® Xeon® Scalable (Ice Lake) processors that deliver an all-core turbo frequency of 3.5 GHz to provide consistent computing performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

i4r instance types

Instance type

vCPUs

Memory (GiB)

Local storage

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.i4r.4xlarge

16

128

1 * 959 GB (1 * 894 GiB)

8/25

3,000,000

100,000

6

ecs.i4r.8xlarge

32

256

1 * 1,919 GB (1 * 1,788 GiB)

16/25

6,000,000

150,000

8

ecs.i4r.16xlarge

64

512

2 * 1,919 GB (2 * 1,788 GiB)

32/none

12,000,000

300,000

16

ecs.i4r.32xlarge

128

1,024

4 * 1,919 GB (4 * 1,788 GiB)

64/none

24,000,000

600,000

32

i4p, performance-enhanced instance family with local SSDs

  • Introduction: This instance family uses the Intel® Second-generation Optane persistent memory (BPS) to provide ultra-high-performance local disks. For information about how to initialize local disks, see the Configure persistent memory as a local disk section of the "Configure the usage mode of persistent memory" topic.

  • Use cases:

    • Gene sequencing applications. For more information, see Case description.

    • On-disk key-value (KV) databases, such as RocksDB and ClickHouse.

    • OLTP and high-performance relational databases for write-ahead log (WAL) optimization.

    • NoSQL databases, such as Cassandra, MongoDB, and HBase.

    • Search scenarios that use solutions such as Elasticsearch.

    • Other I/O-intensive applications that frequently write data to disks, such as message middleware and containers.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4.

    • Uses the third-generation Intel® Xeon® Scalable (Ice Lake) processors that deliver a base frequency of 2.7 GHz and an all-core turbo frequency of 3.2 GHz to provide consistent computing performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

i4p instance types

Instance type

vCPUs

Memory (GiB)

Persistent memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.i4p.2xlarge

8

32

1 * 126

5/10

1,600,000

50,000/up to 110,000

3/6

ecs.i4p.4xlarge

16

64

2 * 126

10/25

3,000,000

80,000/up to 110,000

5/6

ecs.i4p.6xlarge

24

96

3 * 126

12/25

4,500,000

110,000/none

6/none

ecs.i4p.8xlarge

32

128

4 * 126

16/25

6,000,000

150,000/none

8/none

ecs.i4p.16xlarge

64

256

1 * 1008

32/none

12,000,000

300,000/none

16/none

ecs.i4p.32xlarge

128

512

2 * 1008

64/none

24,000,000

600,000/none

32/none

i3g, instance family with local SSDs

  • Introduction: This instance family is equipped with high-performance local NVMe SSDs that deliver high IOPS, high I/O throughput, and low latency.

  • Use cases: OLTP and high-performance relational databases, NoSQL databases such as Cassandra, MongoDB, and HBase, and search scenarios that use solutions such as Elasticsearch.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4, which is designed for high-performance databases.

    • Uses 2.5 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processors that deliver a turbo frequency of 3.2 GHz to provide consistent computing performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

i3g instance types

Instance type

vCPUs

Memory (GiB)

Local storage

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.i3g.2xlarge

8

32

1 * 479 GB (1 * 447 GiB)

3/10

1,750,000

52,500

2

ecs.i3g.4xlarge

16

64

1 * 959 GB (1 * 894 GiB)

5/10

3,500,000

84,000

3

ecs.i3g.8xlarge

32

128

2 * 959 GB (2 * 894 GiB)

12/none

7,000,000

157,500

5

ecs.i3g.13xlarge

52

192

3 * 959 GB (3 * 894 GiB)

16/none

12,000,000

252,000

8

ecs.i3g.26xlarge

104

384

6 * 959 GB (6 * 894 GiB)

32/none

24,000,000

500,000

16

Note

This instance family supports only Linux images. When you create an instance of this instance family, you must select a Linux image.

i3, instance family with local SSDs

  • Introduction: This instance family is equipped with high-performance local NVMe SSDs that deliver high IOPS, high I/O throughput, and low latency, and allows damaged disks to be isolated online.

  • Use cases: OLTP and high-performance relational databases, NoSQL databases such as Cassandra and MongoDB, and search scenarios that use solutions such as Elasticsearch.

  • Compute:

    • Uses 2.5 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processors that deliver a turbo frequency of 3.2 GHz to provide consistent computing performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

i3 instance types

Instance type

vCPUs

Memory (GiB)

Local storage

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.i3.xlarge

4

32

1 * 959 GB (1 * 894 GiB)

1.5/10

1,000,000

40,000

1.5

ecs.i3.2xlarge

8

64

1 * 1,919 GB (1 * 1,788 GiB)

2.5/10

1,600,000

50,000

2

ecs.i3.4xlarge

16

128

2 * 1,919 GB (2 * 1,788 GiB)

5/10

3,000,000

80,000

3

ecs.i3.8xlarge

32

256

4 * 1,919 GB (4 * 1,788 GiB)

10/none

6,000,000

150,000

5

ecs.i3.13xlarge

52

384

6 * 1,919 GB (6 * 1,788 GiB)

16/none

9,000,000

240,000

8

ecs.i3.26xlarge

104

768

12 * 1,919 GB (12 * 1,788 GiB)

32/none

24,000,000

480,000

16

Note

This instance family supports only Linux images. When you create an instance of this instance family, you must select a Linux image.

i2, instance family with local SSDs

  • Introduction: This instance family is equipped with high-performance local NVMe SSDs that deliver high IOPS, high I/O throughput, and low latency.

  • Use cases: OLTP and high-performance relational databases, NoSQL databases such as Cassandra, MongoDB, and HBase, and search scenarios that use solutions such as Elasticsearch.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:8, which is designed for high-performance databases.

    • Uses 2.5 GHz Intel® Xeon® Platinum 8163 (Skylake) processors.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports standard SSDs and ultra disks.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

i2 instance types

Instance type

vCPUs

Memory (GiB)

Local storage

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (pps)

Disk bandwidth (Gbit/s)

ecs.i2.xlarge

4

32

1 * 959 GB (1 * 894 GiB)

1

500,000

Up to 16

ecs.i2.2xlarge

8

64

1 * 1,919 GB (1 * 1,788 GiB)

2

1,000,000

Up to 16

ecs.i2.4xlarge

16

128

2 * 1,919 GB (2 * 1,788 GiB)

3

1,500,000

Up to 16

ecs.i2.8xlarge

32

256

4 * 1,919 GB (4 * 1,788 GiB)

6

2,000,000

Up to 16

ecs.i2.16xlarge

64

512

8 * 1,919 GB (8 * 1,788 GiB)

10

4,000,000

Up to 16

i2g, instance family with local SSDs

  • Introduction: This instance family is equipped with high-performance local NVMe SSDs that deliver high IOPS, high I/O throughput, and low latency.

  • Use cases: OLTP and high-performance relational databases, NoSQL databases such as Cassandra, MongoDB, and HBase, and search scenarios that use solutions such as Elasticsearch.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4, which is designed for high-performance databases.

    • Uses 2.5 GHz Intel® Xeon® Platinum 8163 (Skylake) processors.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports standard SSDs and ultra disks.

  • Network:

    • Supports only IPv4.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

i2g instance types

Instance type

vCPUs

Memory (GiB)

Local storage

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (pps)

ecs.i2g.2xlarge

8

32

1 * 959 GB (1 * 894 GiB)

2

1,000,000

ecs.i2g.4xlarge

16

64

1 * 1,919 GB (1 * 1,788 GiB)

3

1,500,000

ecs.i2g.8xlarge

32

128

2 * 1,919 GB (2 * 1,788 GiB)

6

2,000,000

ecs.i2g.16xlarge

64

256

4 * 1,919 GB (4 * 1,788 GiB)

10

4,000,000

i2ne, instance family with local SSDs

  • Introduction: This instance family is equipped with high-performance local NVMe SSDs that deliver high IOPS, high I/O throughput, and low latency.

  • Use cases: OLTP and high-performance relational databases, NoSQL databases such as Cassandra, MongoDB, and HBase, and search scenarios that use solutions such as Elasticsearch.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:8, which is designed for high-performance databases.

    • Uses 2.5 GHz Intel® Xeon® Platinum 8163 (Skylake) processors.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports standard SSDs and ultra disks.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

    • Provides a network bandwidth of up to 20 Gbit/s.

i2ne instance types

Instance type

vCPUs

Memory (GiB)

Local storage

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (pps)

Disk bandwidth (Gbit/s)

ecs.i2ne.xlarge

4

32

1 * 959 GB (1 * 894 GiB)

1.5

500,000

Up to 16

ecs.i2ne.2xlarge

8

64

1 * 1,919 GB (1 * 1,788 GiB)

2.5

1,000,000

Up to 16

ecs.i2ne.4xlarge

16

128

2 * 1,919 GB (2 * 1,788 GiB)

5

1,500,000

Up to 16

ecs.i2ne.8xlarge

32

256

4 * 1,919 GB (4 * 1,788 GiB)

10

2,000,000

Up to 16

ecs.i2ne.16xlarge

64

512

8 * 1,919 GB (8 * 1,788 GiB)

20

4,000,000

Up to 16

ecs.i2ne.20xlarge

80

704

10 * 1,919 GB (10 * 1,788 GiB)

25

4,500,000

Up to 16

i2gne, instance family with local SSDs

  • Introduction: This instance family is equipped with high-performance local NVMe SSDs that deliver high IOPS, high I/O throughput, and low latency.

  • Use cases: OLTP and high-performance relational databases, NoSQL databases such as Cassandra, MongoDB, and HBase, and search scenarios that use solutions such as Elasticsearch.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4, which is designed for high-performance databases.

    • Uses 2.5 GHz Intel® Xeon® Platinum 8163 (Skylake) processors.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports standard SSDs and ultra disks.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

    • Provides a network bandwidth of up to 20 Gbit/s.

i2gne instance types

Instance type

vCPUs

Memory (GiB)

Local storage

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (pps)

ecs.i2gne.2xlarge

8

32

1 * 959 GB (1 * 894 GiB)

2.5

1,000,000

ecs.i2gne.4xlarge

16

64

1 * 1,919 GB (1 * 1,788 GiB)

5

1,500,000

ecs.i2gne.8xlarge

32

128

2 * 1,919 GB (2 * 1,788 GiB)

10

2,000,000

ecs.i2gne.16xlarge

64

256

4 * 1,919 GB (4 * 1,788 GiB)

20

4,000,000

hfc9i, compute-optimized instance family with high clock speeds

  • Introduction: This instance family uses Alibaba Cloud's new CIPU architecture with performance cores (P-core) of Intel ®Xeon ®6 processor. It provides stable computing power output, a more powerful I/O engine, and chip-level security hardening.

  • Scenarios: high network packet sending and receiving scenarios, data analysis, Batch Compute, video encoding, massively multiplayer online game (MMO) frontend, and high-performance scientific and engineering applications.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:2.

    • Uses Intel ®Xeon ®Granite Rapids processors, with a clock frequency of 3.4 GHz and an all-core turbo frequency of 3.8 GHz for stable computing performance.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports ESSDs, ESSD AutoPL disks, and Regional ESSDs. For information about disks, see Overview of Block Storage.

    • These instances support the NVMe protocol. For more information, see Overview of the NVMe protocol.

    • For small instance types, disk IOPS and disk bandwidth are burstable. The storage I/O performance of an instance corresponds to its instance type. A larger instance type provides higher storage I/O performance. For more information, see Storage I/O performance.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: supports the virtual Trusted Platform Module (vTPM) feature. For more information, see Overview of trusted computing capabilities.

hfc9i instance types

Instance type

vCPU

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.hfc9i.large

2

4

2.5/15

1,000,000

500,000

2

3

6

6

20,000/burstable up to 200,000

2/12

ecs.hfc9i.xlarge

4

8

4/15

1,200,000

500,000

4

4

15

15

50,000/burstable up to 200,000

2.5/12

ecs.hfc9i.2xlarge

8

16

8/15

1,600,000

500,000

8

4

15

15

60,000/burstable up to 200,000

4/12

ecs.hfc9i.3xlarge

12

24

10/15

2,400,000

500,000

12

8

15

15

80,000/burstable up to 200,000

5/12

ecs.hfc9i.4xlarge

16

32

16/25

3,000,000

500,000

16

8

30

30

100,000/burstable up to 200,000

6/12

ecs.hfc9i.6xlarge

24

48

18/25

4,500,000

500,000

24

8

30

30

120,000/burstable up to 200,000

8/12

ecs.hfc9i.8xlarge

32

64

20/32

6,000,000

800,000

32

8

30

30

200,000/burstable up to 250,000

12/15

ecs.hfc9i.12xlarge

48

96

25/32

9,000,000

1,000,000

48

8

30

30

250,000/none

15/none

ecs.hfc9i.16xlarge

64

128

36/none

12,000,000

2,000,000

64

8

30

30

360,000/none

20/none

ecs.hfc9i.24xlarge

96

192

48/none

16,000,000

3,000,000

64

15

50

50

400,000/None

24/none

ecs.hfc9i.36xlarge

144

384

64/none

20,000,000

4,000,000

64

15

50

50

500,000/none

32/none

hfc8i, compute-optimized instance family with high clock speeds

  • Introduction: This instance family uses the innovative CIPU architecture developed by Alibaba Cloud to provide stable computing power, a more robust I/O engine, and chip-level security hardening.

  • Scenarios:

    • Large volumes of packets are received and transmitted, such as live commenting on videos and telecom data forwarding

    • High-performance frontend server clusters

    • Frontend servers of massively multiplayer online (MMO) games

    • Data analytics, batch processing, and video encoding

    • High-performance scientific and engineering applications

  • Compute:

    • Offers a CPU-to-memory ratio of 1:2.

    • Uses the fourth-generation Intel ®Xeon ®Scalable processors (Sapphire Rapids) with a base frequency of 3.3 GHz and an all-core turbo frequency of 3.9 GHz for stable computing performance.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

    • Is compatible with specific operating systems. For more information, see OS compatibility for Intel instance types.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • Provides burstable network bandwidth for low-specification instances.

    • Provides high network performance based on large computing capacity.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

hfc8i instance types

Instance type

vCPU

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.hfc8i.large

2

4

2.5/15

1,200,000

300,000

2

3

6

6

30,000/burstable up to 200,000

3/12

ecs.hfc8i.xlarge

4

8

4/15

1,400,000

300,000

4

4

15

15

50,000/burstable up to 200,000

4/12

ecs.hfc8i.2xlarge

8

16

8/15

1,800,000

300,000

8

4

15

15

60,000/burstable up to 200,000

6/12

ecs.hfc8i.3xlarge

12

24

10/15

2,800,000

300,000

12

8

15

15

90,000/burstable up to 200,000

8/12

ecs.hfc8i.4xlarge

16

32

16/25

3,600,000

500,000

16

8

30

30

120,000/burstable up to 200,000

10/12

ecs.hfc8i.6xlarge

24

48

18/25

5,500,000

800,000

24

8

30

30

200,000/none

12/none

ecs.hfc8i.8xlarge

32

64

32/none

7,500,000

1,000,000

32

8

30

30

250,000/none

16/none

ecs.hfc8i.16xlarge

64

128

64/none

15,000,000

2,000,000

64

8

30

30

450,000 /None

32/none

ecs.hfc8i.32xlarge

128

256

100/none

30,000,000

4,000,000

64

15

50

50

900,000/none

64/none

hfc7, compute-optimized instance family with high clock speeds

  • Introduction: This instance family offloads a large number of virtualization features to dedicated hardware by using the third-generation SHENLONG architecture to provide predictable and consistent ultra-high performance and reduce virtualization overheads.

  • Scenarios:

    • Large volumes of packets are received and transmitted, such as live commenting on videos and telecom data forwarding

    • High-performance frontend server clusters

    • Frontend servers of massively multiplayer online (MMO) games

    • Data analytics, batch processing, and video encoding

    • High-performance scientific and engineering applications

  • Compute:

    • Offers a CPU-to-memory ratio of 1:2.

    • Uses Intel® Xeon® Cooper Lake processors that deliver an all-core turbo frequency of 3.8 GHz and a clock speed of at least 3.3 GHz to provide consistent computing performance.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Provides ultra-high packet forwarding rates.

    • Provides high network performance based on large computing capacity.

hfc7 instance types

Instance type

vCPU

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.hfc7.large

2

4

1.2/10

900,000

250,000

2

2

6

6

20,000

1

ecs.hfc7.xlarge

4

8

2/10

1,000,000

250,000

4

3

15

15

30,000

1.5

ecs.hfc7.2xlarge

8

16

3/10

1,600,000

250,000

8

4

15

15

45,000

2

ecs.hfc7.3xlarge

12

24

4.5/10

2,000,000

250,000

8

6

15

15

60,000

2.5

ecs.hfc7.4xlarge

16

32

6/10

2,500,000

300,000

8

8

30

30

75,000

3

ecs.hfc7.6xlarge

24

48

8/10

3,000,000

450,000

12

8

30

30

90,000

4

ecs.hfc7.8xlarge

32

64

10/none

4,000,000

600,000

16

8

30

30

105,000

5

ecs.hfc7.12xlarge

48

96

16/none

6,000,000

1,000,000

24

8

30

30

150,000

8

ecs.hfc7.24xlarge

96

192

32/none

12,000,000

1,800,000

32

15

30

30

300,000

16

hfc6, compute-optimized instance family with high clock speeds

  • Introduction: This instance family offloads a large number of virtualization features to dedicated hardware by using the SHENLONG architecture to provide predictable and consistent ultra-high performance and reduce virtualization overheads.

  • Scenarios:

    • Large volumes of packets are received and transmitted, such as live commenting on videos and telecom data forwarding

    • Web frontend servers

    • Frontend servers of massively multiplayer online (MMO) games

    • Data analytics, batch processing, and video encoding

    • High-performance scientific and engineering applications

  • Compute:

    • Offers a CPU-to-memory ratio of 1:2.

    • Uses 3.1 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processors that deliver a turbo frequency of 3.5 GHz to provide consistent computing performance.

      Note

      The processors used by this instance family have a clock speed of 3.1 GHz. However, the Intel System Studio (ISS) feature may cause a lower clock speed to be displayed. Alibaba Cloud is working on this issue. This issue does not affect the actual clock speeds of your instances.

      You can separately run the following commands to use the turbostat tool to view the actual clock speeds:

      yum install kernel-tools
      turbostat
    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Provides ultra-high packet forwarding rates.

    • Provides high network performance based on large computing capacity.

hfc6 instance types

Instance type

vCPU

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.hfc6.large

2

4

1/3

300,000

35,000

2

2

6

1

10,000

1

ecs.hfc6.xlarge

4

8

1.5/5

500,000

70,000

4

3

10

1

20,000

1.5

ecs.hfc6.2xlarge

8

16

2.5/8

800,000

150,000

8

4

10

1

25,000

2

ecs.hfc6.3xlarge

12

24

4/10

900,000

220,000

8

6

10

1

30,000

2.5

ecs.hfc6.4xlarge

16

32

5/10

1,000,000

300,000

8

8

20

1

40,000

3

ecs.hfc6.6xlarge

24

48

7.5/10

1,500,000

450,000

12

8

20

1

50,000

4

ecs.hfc6.8xlarge

32

64

10/none

2,000,000

600,000

16

8

20

1

60,000

5

ecs.hfc6.10xlarge

40

96

12.5/none

3,000,000

1,000,000

32

7

20

1

100,000

8

ecs.hfc6.16xlarge

64

128

20/none

4,000,000

1,200,000

32

8

20

1

120,000

10

ecs.hfc6.20xlarge

80

192

25/none

6,000,000

1,800,000

32

15

20

1

200,000

16

hfg9i, general-purpose instance family with high clock speeds

  • Introduction: This instance family uses Alibaba Cloud's new CIPU architecture with P-core (performance cores) of Intel ®Xeon ®6 processor. It provides stable computing power output, a more powerful I/O engine, and chip-level security hardening.

  • Scenarios: high network packet sending and receiving scenarios, data analysis, Batch Compute, and video encoding, massively multiplayer online game (MMO) frontend, high-performance scientific and engineering applications, and medium and large databases.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4.

    • Uses Intel ®Xeon ®Granite Rapids processors, with a clock frequency of 3.4 GHz and an all-core turbo frequency of 3.8 GHz for stable computing performance.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports ESSDs, ESSD AutoPL disks, and Regional ESSDs. For information about disks, see Overview of Block Storage.

    • These instances support the NVMe protocol. For more information, see Overview of the NVMe protocol.

    • For small instance types, disk IOPS and disk bandwidth are burstable. The storage I/O performance of an instance corresponds to its instance type. A larger instance type provides higher storage I/O performance. For more information, see Storage I/O performance.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

hfg9i instance types

Instance type

vCPU

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.hfg9i.large

2

8

2.5/15

1,000,000

500,000

2

3

6

6

20,000/burstable up to 200,000

2/12

ecs.hfg9i.xlarge

4

16

4/15

1,200,000

500,000

4

4

15

15

50,000/burstable up to 200,000

2.5/12

ecs.hfg9i.2xlarge

8

32

8/15

1,600,000

500,000

8

4

15

15

60,000/burstable up to 200,000

4/12

ecs.hfg9i.3xlarge

12

48

10/15

2,400,000

500,000

12

8

15

15

80,000/burstable up to 200,000

5/12

ecs.hfg9i.4xlarge

16

64

16/25

3,000,000

500,000

16

8

30

30

100,000/burstable up to 200,000

6/12

ecs.hfg9i.6xlarge

24

96

18/25

4,500,000

500,000

24

8

30

30

120,000/burstable up to 200,000

8/12

ecs.hfg9i.8xlarge

32

128

20/32

6,000,000

800,000

32

8

30

30

200,000 /burstable up to 250,000

12/15

ecs.hfg9i.12xlarge

48

192

25/32

9,000,000

1,000,000

48

8

30

30

250,000/none

15/none

ecs.hfg9i.16xlarge

64

256

36/none

12,000,000

2,000,000

64

8

30

30

360,000/none

20/none

ecs.hfg9i.24xlarge

96

384

48/none

16,000,000

3,000,000

64

15

50

50

400,000 /None

24/none

ecs.hfg9i.36xlarge

144

768

64/none

20,000,000

4,000,000

64

15

50

50

500,000/none

32/none

hfg8i, general-purpose instance family with high clock speeds

  • Introduction: This instance family uses the innovative CIPU architecture developed by Alibaba Cloud to provide stable computing power, a more robust I/O engine, and chip-level security hardening.

  • Scenarios:

    • Large volumes of packets are received and transmitted, such as live commenting on videos and telecom data forwarding

    • High-performance frontend server clusters

    • High-performance frontend server clusters

    • Data analytics, batch processing, and video encoding

    • High-performance scientific and engineering applications

    • Medium-sized database systems

    • Enterprise-level applications of various types and sizes

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4.

    • Uses the fourth-generation Intel ®Xeon ®Scalable processors (Sapphire Rapids) with a base frequency of 3.3 GHz and an all-core turbo frequency of 3.9 GHz for stable computing performance.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

    • Is compatible with specific operating systems. For more information, see OS compatibility for Intel instance types.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • Provides high storage I/O performance based on large computing capacity. For more information, see Storage I/O performance.

    • Provides high network performance based on large computing capacity.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

hfg8i instance types

Instance type

vCPU

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Attachable disks

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.hfg8i.large

2

8

2.5/15

1,200,000

300,000

2

3

6

6

16

30,000/burstable up to 200,000

3/12

ecs.hfg8i.xlarge

4

16

4/15

1,400,000

300,000

4

4

15

15

16

50,000/burstable up to 200,000

4/12

ecs.hfg8i.2xlarge

8

32

8/15

1,800,000

300,000

8

4

15

15

16

60,000/burstable up to 200,000

6/12

ecs.hfg8i.3xlarge

12

48

10/15

2,800,000

300,000

12

8

15

15

16

90,000/burstable up to 200,000

8/12

ecs.hfg8i.4xlarge

16

64

16/25

3,600,000

500,000

16

8

30

30

16

120,000/burstable up to 200,000

10/12

ecs.hfg8i.6xlarge

24

96

18/25

5,500,000

800,000

24

8

30

30

24

200,000/none

12/none

ecs.hfg8i.8xlarge

32

128

32/none

7,500,000

1,000,000

32

8

30

30

32

250,000/none

16/none

ecs.hfg8i.16xlarge

64

256

64/none

15,000,000

2,000,000

64

8

30

30

32

450,000 /None

32/none

ecs.hfg8i.32xlarge

128

512

100/none

30,000,000

4,000,000

64

15

50

50

64

900,000/none

64/none

hfg7, general-purpose instance family with high clock speeds

  • Introduction: This instance family offloads a large number of virtualization features to dedicated hardware by using the third-generation SHENLONG architecture to provide predictable and consistent ultra-high performance and reduce virtualization overheads.

  • Scenarios:

    • Large volumes of packets are received and transmitted, such as live commenting on videos and telecom data forwarding

    • Enterprise-grade applications of various types and sizes

    • Game servers

    • Small and medium-sized database systems, caches, and search clusters

    • High-performance scientific computing

    • Video encoding applications

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4.

    • Uses Intel® Xeon® Cooper Lake processors that deliver an all-core turbo frequency of 3.8 GHz and a clock speed of at least 3.3 GHz to provide consistent computing performance.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Provides ultra-high packet forwarding rates.

    • Provides high network performance based on large computing capacity.

hfg7 instance types

Instance type

vCPU

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.hfg7.large

2

8

1.2/10

900,000

250,000

2

2

6

6

20,000

1

ecs.hfg7.xlarge

4

16

2/10

1,000,000

250,000

4

3

15

15

30,000

1.5

ecs.hfg7.2xlarge

8

32

3/10

1,600,000

250,000

8

4

15

15

45,000

2

ecs.hfg7.3xlarge

12

48

4.5/10

2,000,000

250,000

8

6

15

15

60,000

2.5

ecs.hfg7.4xlarge

16

64

6/10

2,500,000

300,000

8

8

30

30

75,000

3

ecs.hfg7.6xlarge

24

96

8/10

3,000,000

450,000

12

8

30

30

90,000

4

ecs.hfg7.8xlarge

32

128

10/none

4,000,000

600,000

16

8

30

30

105,000

5

ecs.hfg7.12xlarge

48

192

16/none

6,000,000

1,000,000

24

8

30

30

150,000

8

ecs.hfg7.24xlarge

96

384

32/none

12,000,000

1,800,000

32

15

30

30

300,000

16

hfg6, general-purpose instance family with high clock speeds

  • Introduction: This instance family offloads a large number of virtualization features to dedicated hardware by using the SHENLONG architecture to provide predictable and consistent ultra-high performance and reduce virtualization overheads.

  • Scenarios:

    • Large volumes of packets are received and transmitted, such as live commenting on videos and telecom data forwarding

    • Enterprise-grade applications of various types and sizes

    • Websites and application servers

    • Game servers

    • Small and medium-sized database systems, caches, and search clusters

    • Data analytics and computing

    • Computing clusters and memory-intensive data processing

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4.

    • Uses 3.1 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processors that deliver a turbo frequency of 3.5 GHz to provide consistent computing performance.

      Note

      The processors used by this instance family have a clock speed of 3.1 GHz. However, the Intel System Studio (ISS) feature may cause a lower clock speed to be displayed. Alibaba Cloud is working on this issue. This issue does not affect the actual clock speeds of your instances.

      You can separately run the following commands to use the turbostat tool to view the actual clock speeds:

      yum install kernel-tools
      turbostat
    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Provides ultra-high packet forwarding rates.

    • Provides high network performance based on large computing capacity.

hfg6 instance types

Instance type

vCPU

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.hfg6.large

2

8

1/3

300,000

35,000

2

2

6

1

10,000

1

ecs.hfg6.xlarge

4

16

1.5/5

500,000

70,000

4

3

10

1

20,000

1.5

ecs.hfg6.2xlarge

8

32

2.5/8

800,000

150,000

8

4

10

1

25,000

2

ecs.hfg6.3xlarge

12

48

4/10

900,000

220,000

8

6

10

1

30,000

2.5

ecs.hfg6.4xlarge

16

64

5/10

1,000,000

300,000

8

8

20

1

40,000

3

ecs.hfg6.6xlarge

24

96

7.5/10

1,500,000

450,000

12

8

20

1

50,000

4

ecs.hfg6.8xlarge

32

128

10/none

2,000,000

600,000

16

8

20

1

60,000

5

ecs.hfg6.10xlarge

40

192

12.5/none

3,000,000

1,000,000

32

7

20

1

100,000

8

ecs.hfg6.16xlarge

64

256

20/none

4,000,000

1,200,000

32

8

20

1

120,000

10

ecs.hfg6.20xlarge

80

384

25/none

6,000,000

1,800,000

32

15

20

1

200,000

16

hfr9i, memory-optimized instance family with high clock speeds

  • Introduction: This instance family uses Alibaba Cloud's new CIPU architecture with P-core (performance cores) of Intel ®Xeon ®6 processor. It provides stable computing power output, a more powerful I/O engine, and chip-level security hardening.

  • Scenarios: high-network packet sending and receiving scenarios, data analysis and mining, distributed memory caching, high-performance databases, in-memory databases, high-performance scientific and engineering applications, Hadoop and Spark clusters, and other enterprise applications that require large amounts of memory.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:8.

    • Uses Intel ®Xeon ®Granite Rapids processors, with a clock frequency of 3.4 GHz and an all-core turbo frequency of 3.8 GHz for stable computing performance.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports ESSDs, ESSD AutoPL disks, and Regional ESSDs. For information about disks, see Overview of Block Storage.

    • These instances support the NVMe protocol. For more information, see Overview of the NVMe protocol.

    • For small instance types, disk IOPS and disk bandwidth are burstable. The storage I/O performance of an instance corresponds to its instance type. A larger instance type provides higher storage I/O performance. For more information, see Storage I/O performance.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

hfr9i instance types

Instance type

vCPU

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.hfr9i.large

2

16

2.5/15

1,000,000

500,000

2

3

6

6

20,000/burstable up to 200,000

2/12

ecs.hfr9i.xlarge

4

32

4/15

1,200,000

500,000

4

4

15

15

50,000/burstable up to 200,000

2.5/12

ecs.hfr9i.2xlarge

8

64

8/15

1,600,000

500,000

8

4

15

15

60,000/burstable up to 200,000

4/12

ecs.hfr9i.3xlarge

12

96

10/15

2,400,000

500,000

12

8

15

15

80,000/burstable up to 200,000

5/12

ecs.hfr9i.4xlarge

16

128

16/25

3,000,000

500,000

16

8

30

30

100,000/burstable up to 200,000

6/12

ecs.hfr9i.6xlarge

24

192

18/25

4,500,000

500,000

24

8

30

30

120,000/burstable up to 200,000

8/12

ecs.hfr9i.8xlarge

32

256

20/32

6,000,000

800,000

32

8

30

30

200,000/burstable up to 250,000

12/15

ecs.hfr9i.12xlarge

48

384

25/32

9,000,000

1,000,000

48

8

30

30

250,000/none

15/none

ecs.hfr9i.16xlarge

64

512

36/none

12,000,000

2,000,000

64

8

30

30

360,000/none

20/none

ecs.hfr9i.24xlarge

96

768

48/none

16,000,000

3,000,000

64

15

50

50

400,000/None

24/none

ecs.hfr9i.36xlarge

144

1536

64/none

20,000,000

4,000,000

64

15

50

50

500,000/none

32/none

hfr8i, memory-optimized instance family with high clock speeds

  • Introduction: This instance family uses the innovative CIPU architecture developed by Alibaba Cloud to provide stable computing power, a more robust I/O engine, and chip-level security hardening.

  • Scenarios:

    • Large volumes of packets are received and transmitted, such as live commenting on videos and telecom data forwarding

    • High-performance scientific and engineering applications

    • High-performance databases and in-memory databases

    • Data analytics, data mining, and distributed memory caching

    • Enterprise-level memory-intensive applications such as Hadoop clusters and Spark clusters

  • Compute:

    • Offers a CPU-to-memory ratio of 1:8.

    • Uses the fourth-generation Intel® Xeon® Scalable (Sapphire Rapids) processors that deliver a base frequency of 2.7 GHz and an all-core turbo frequency of 3.2 GHz to provide consistent computing performance.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

    • Is compatible with specific operating systems. For more information, see OS compatibility for Intel instance types.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Provides burstable network bandwidth for low-specification instances.

    • Provides high network performance based on large computing capacity.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

hfr8i instance types

Instance type

vCPU

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Attachable disks

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.hfr8i.large

2

16

2.5/15

1,200,000

300,000

2

3

6

6

16

30,000/burstable up to 200,000

3/12

ecs.hfr8i.xlarge

4

32

4/15

1,400,000

300,000

4

4

15

15

16

50,000/burstable up to 200,000

4/12

ecs.hfr8i.2xlarge

8

64

8/15

1,800,000

300,000

8

4

15

15

16

60,000/burstable up to 200,000

6/12

ecs.hfr8i.3xlarge

12

96

10/15

2,800,000

300,000

12

8

15

15

16

90,000/burstable up to 200,000

8/12

ecs.hfr8i.4xlarge

16

128

16/25

3,600,000

500,000

16

8

30

30

16

120,000/burstable up to 200,000

10/12

ecs.hfr8i.6xlarge

24

192

18/25

5,500,000

800,000

24

8

30

30

24

200,000/none

12/none

ecs.hfr8i.8xlarge

32

256

32/none

7,500,000

1,000,000

32

8

30

30

32

250,000/none

16/none

ecs.hfr8i.16xlarge

64

512

64/none

15,000,000

2,000,000

64

8

30

30

32

450,000/None

32/none

ecs.hfr8i.32xlarge

128

1024

100/none

30,000,000

4,000,000

64

15

50

50

64

900,000/none

64/none

hfr7, memory-optimized instance family with high clock speeds

  • Introduction: This instance family offloads a large number of virtualization features to dedicated hardware by using the third-generation SHENLONG architecture to provide predictable and consistent ultra-high performance and reduce virtualization overheads.

  • Scenarios:

    • Large volumes of packets are received and transmitted, such as live commenting and telecom data forwarding

    • High-performance databases and in-memory databases

    • Data analytics, data mining, and distributed memory caching

    • Enterprise-level memory-intensive applications such as Hadoop clusters and Spark clusters

  • Compute:

    • Offers a CPU-to-memory ratio of 1:8.

    • Uses Intel® Xeon® Cooper Lake processors that deliver an all-core turbo frequency of 3.8 GHz and a clock speed of at least 3.3 GHz to provide consistent computing performance.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Provides ultra-high packet forwarding rates.

    • Provides high network performance based on large computing capacity.

hfr7 instance types

Instance type

vCPU

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.hfr7.large

2

16

1.2/10

900,000

250,000

2

2

6

6

20,000

1

ecs.hfr7.xlarge

4

32

2/10

1,000,000

250,000

4

3

15

15

30,000

1.5

ecs.hfr7.2xlarge

8

64

3/10

1,600,000

250,000

8

4

15

15

45,000

2

ecs.hfr7.3xlarge

12

96

4.5/10

2,000,000

250,000

8

6

15

15

60,000

2.5

ecs.hfr7.4xlarge

16

128

6/10

2,500,000

300,000

8

8

30

30

75,000

3

ecs.hfr7.6xlarge

24

192

8/10

3,000,000

450,000

12

8

30

30

90,000

4

ecs.hfr7.8xlarge

32

256

10/none

4,000,000

600,000

16

8

30

30

105,000

5

ecs.hfr7.12xlarge

48

384

16/none

6,000,000

1,000,000

24

8

30

30

150,000

8

ecs.hfr7.24xlarge

96

768

32/none

12,000,000

1,800,000

32

15

30

30

300,000

16

hfr6, memory-optimized instance family with high clock speeds

  • Introduction: This instance family offloads a large number of virtualization features to dedicated hardware by using the SHENLONG architecture to provide predictable and consistent ultra-high performance and reduce virtualization overheads.

  • Scenarios:

    • Large volumes of packets are received and transmitted, such as live commenting on videos and telecom data forwarding

    • High-performance databases and in-memory databases

    • Data analytics, data mining, and distributed memory caching

    • Enterprise-level memory-intensive applications such as Hadoop clusters and Spark clusters

  • Compute:

    • Offers a CPU-to-memory ratio of 1:8.

    • Uses 3.1 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processors that deliver a turbo frequency of 3.5 GHz to provide consistent computing performance.

      Note

      The processors used by this instance family have a clock speed of 3.1 GHz. However, the Intel System Studio (ISS) feature may cause a lower clock speed to be displayed. Alibaba Cloud is working on this issue. This issue does not affect the actual clock speeds of your instances.

      You can separately run the following commands to use the turbostat tool to view the actual clock speeds:

      yum install kernel-tools
      turbostat
    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Provides ultra-high packet forwarding rates.

    • Provides high network performance based on large computing capacity.

hfr6 instance types

Instance type

vCPU

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.hfr6.large

2

16

1/3

300,000

35,000

2

2

6

1

10,000

1

ecs.hfr6.xlarge

4

32

1.5/5

500,000

70,000

4

3

10

1

20,000

1.5

ecs.hfr6.2xlarge

8

64

2.5/8

800,000

150,000

8

4

10

1

25,000

2

ecs.hfr6.3xlarge

12

96

4/10

900,000

220,000

8

6

10

1

30,000

2.5

ecs.hfr6.4xlarge

16

128

5/10

1,000,000

300,000

8

8

20

1

40,000

3

ecs.hfr6.6xlarge

24

192

7.5/10

1,500,000

450,000

12

8

20

1

50,000

4

ecs.hfr6.8xlarge

32

256

10/none

2,000,000

600,000

16

8

20

1

60,000

5

ecs.hfr6.10xlarge

40

384

12.5/none

3,000,000

1,000,000

32

7

20

1

100,000

8

ecs.hfr6.16xlarge

64

512

20/none

4,000,000

1,200,000

32

8

20

1

120,000

10

ecs.hfr6.20xlarge

80

768

25/none

6,000,000

1,800,000

32

15

20

1

200,000

16

hfc5, compute-optimized instance family with high clock speeds

  • Scenarios: High-performance frontend servers, high-performance scientific and engineering applications, MMO games, and video encoding.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:2.

    • Uses processors of 3.1 GHz Intel ®Xeon ®Gold 6149(Skylake) or 8269CY(Cascade Lake) to provide stable computing performance.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports standard SSDs and ultra disks. For information about disks, see Overview of Block Storage.

  • Network:

    • Supports only IPv4.

    • Provides high network performance based on large computing capacity.

hfc5 instance types

Instance type

vCPU

Memory (GiB)

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (pps)

NIC queues

ENIs

Private IPv4 addresses per ENI

ecs.hfc5.large

2

4

1

300,000

2

2

6

ecs.hfc5.xlarge

4

8

1.5

500,000

2

3

10

ecs.hfc5.2xlarge

8

16

2

1,000,000

4

4

10

ecs.hfc5.3xlarge

12

24

2.5

1,300,000

4

6

10

ecs.hfc5.4xlarge

16

32

3

1,600,000

4

8

20

ecs.hfc5.6xlarge

24

48

4.5

2,000,000

6

8

20

ecs.hfc5.8xlarge

32

64

6

2,500,000

8

8

20

hfg5, general-purpose instance family with high clock speeds

  • Scenarios: High-performance frontend servers, high-performance scientific and engineering applications, MMO games, and video encoding.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4 (excluding the instance type with 56 vCPUs).

    • Uses 3.1 GHz Intel® Xeon® Gold 6149 (Skylake) processors to provide consistent computing performance.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports standard SSDs and ultra disks. For information about disks, see Overview of Block Storage.

  • Network:

    • Supports only IPv4.

    • Provides high network performance based on large computing capacity.

hfg5 instance types

Instance type

vCPU

Memory (GiB)

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (pps)

NIC queues

ENIs

Private IPv4 addresses per ENI

ecs.hfg5.large

2

8

1

300,000

2

2

6

ecs.hfg5.xlarge

4

16

1.5

500,000

2

3

10

ecs.hfg5.2xlarge

8

32

2

1,000,000

4

4

10

ecs.hfg5.3xlarge

12

48

2.5

1,300,000

4

6

10

ecs.hfg5.4xlarge

16

64

3

1,600,000

4

8

20

ecs.hfg5.6xlarge

24

96

4.5

2,000,000

6

8

20

ecs.hfg5.8xlarge

32

128

6

2,500,000

8

8

20

ecs.hfg5.14xlarge

56

160

10

4,000,000

14

8

20

g7se, storage-enhanced general-purpose instance family

  • Introduction: This instance family uses the third-generation SHENLONG architecture and Intel Ice Lake processors to improve storage I/O performance.

  • Supported scenarios: I/O-intensive scenarios such as large and medium-sized online transaction processing (OLTP) core databases, large and medium-sized NoSQL databases, search and real-time log analytics, and traditional large enterprise-level commercial software such as SAP.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4.

    • Uses the third-generation Intel® Xeon® Scalable (Ice Lake) processors that deliver a base frequency of 2.9 GHz and an all-core turbo frequency of 3.5 GHz to provide consistent computing performance.

    • Supports Hyper-Threading. By default, Hyper-Threading is enabled. For more information, see Change CPU options.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports the Non-Volatile Memory Express (NVMe) protocol. For more information, see NVMe protocol.

    • Supports Enterprise SSDs (ESSDs), ESSD AutoPL disks, and Regional ESSDs. For information about disks, see Overview of Block Storage.

    • Allows up to 64 data disks to be attached per instance. You can attach up to 16 data disks to an instance when you create the instance. If the instance requires additional data disks, attach more data disks after the instance is created. For more information, see Attach a data disk.

    • Delivers a sequential read/write throughput of up to 64 Gbit/s and up to 1,000,000 IOPS per instance.

    • Provides high storage I/O performance based on large computing capacity. For more information, see Storage I/O performance.

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances provide an ultra-high packet forwarding rate.

    • Provides high network performance based on large computing capacity.

g7se instance types

Instance type

vCPUs

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Maximum attached data disks

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.g7se.large

2

8

1.2/burstable up to 3

450,000

Up to 250,000

2

3

6

6

16

30,000/burstable up to 150,000

3/10

ecs.g7se.xlarge

4

16

2/burstable up to 5

500,000

Up to 250,000

4

4

15

15

16

60,000/burstable up to 150,000

4/10

ecs.g7se.2xlarge

8

32

3/burstable up to 8

800,000

Up to 250,000

8

4

15

15

16

100,000/burstable up to 150,000

6/10

ecs.g7se.3xlarge

12

48

4.5/burstable up to 10

1,200,000

Up to 250,000

8

8

15

15

16

120,000/burstable up to 150,000

8/10

ecs.g7se.4xlarge

16

64

6/burstable up to 10

1,500,000

300,000

8

8

30

30

24

150,000/none

10/none

ecs.g7se.6xlarge

24

96

8/burstable up to 10

2,250,000

450,000

12

8

30

30

24

200,000/none

12/none

ecs.g7se.8xlarge

32

128

10/none

3,000,000

600,000

16

8

30

30

30

300,000/none

16/none

ecs.g7se.16xlarge

64

256

16/none

6,000,000

1,200,000

32

8

30

30

56

500,000/none

32/none

c7se, storage-enhanced compute-optimized instance family

  • Introduction: This instance family uses the third-generation SHENLONG architecture and Intel Ice Lake processors to improve storage I/O performance.

  • Supported scenarios: I/O-intensive scenarios such as large and medium-sized OLTP core databases, large and medium-sized NoSQL databases, search and real-time log analytics, and traditional large enterprise-level commercial software such as SAP.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:2.

    • Uses the third-generation Intel® Xeon® Scalable (Ice Lake) processors that deliver a base frequency of 2.9 GHz and an all-core turbo frequency of 3.5 GHz to provide consistent computing performance.

    • Supports Hyper-Threading. By default, Hyper-Threading is enabled. For more information, see Change CPU options.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports the Non-Volatile Memory Express (NVMe) protocol. For more information, see NVMe protocol.

    • Supports ESSDs, ESSD AutoPL disks, and Regional ESSDs. For information about disks, see Overview of Block Storage.

    • Allows up to 64 data disks to be attached per instance. You can attach up to 16 data disks to an instance when you create the instance. If the instance requires additional data disks, attach more data disks after the instance is created. For more information, see Attach a data disk.

    • Delivers a sequential read/write throughput of up to 64 Gbit/s and up to 1,000,000 IOPS per instance.

    • Provides high storage I/O performance based on large computing capacity. For more information, see Storage I/O performance.

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances provide an ultra-high packet forwarding rate.

    • Provides high network performance based on large computing capacity.

c7se instance types

Instance type

vCPUs

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Maximum attached data disks

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.c7se.large

2

4

1.2/burstable up to 3

450,000

Up to 250,000

2

3

6

6

16

30,000/burstable up to 150,000

3/10

ecs.c7se.xlarge

4

8

2/burstable up to 5

500,000

Up to 250,000

4

4

15

15

16

60,000/burstable up to 150,000

4/10

ecs.c7se.2xlarge

8

16

3/burstable up to 8

800,000

Up to 250,000

8

4

15

15

16

100,000/burstable up to 150,000

6/10

ecs.c7se.3xlarge

12

24

4.5/burstable up to 10

1,200,000

Up to 250,000

8

8

15

15

16

120,000/burstable up to 150,000

8/10

ecs.c7se.4xlarge

16

32

6/burstable up to 10

1,500,000

300,000

8

8

30

30

24

150,000/none

10/none

ecs.c7se.6xlarge

24

48

8/burstable up to 10

2,250,000

450,000

12

8

30

30

24

200,000/none

12/none

ecs.c7se.8xlarge

32

64

10/none

3,000,000

600,000

16

8

30

30

30

300,000/none

16/none

ecs.c7se.16xlarge

64

128

16/none

6,000,000

1,200,000

32

8

30

30

56

500,000/none

32/none

r7se, storage-enhanced memory-optimized instance family

  • Introduction: This instance family uses the third-generation SHENLONG architecture and Intel Ice Lake processors to improve storage I/O performance.

  • Supported scenarios:

    • I/O-intensive scenarios such as large and medium-sized OLTP core databases

    • Large and medium-sized NoSQL databases

    • Search and real-time log analytics

    • Traditional large enterprise-level commercial software such as SAP

    • High-density deployment of containers

  • Compute:

    • Offers a CPU-to-memory ratio of 1:8.

    • Uses the third-generation Intel® Xeon® Scalable (Ice Lake) processors that deliver a base frequency of 2.9 GHz and an all-core turbo frequency of 3.5 GHz to provide consistent computing performance.

    • Supports Hyper-Threading. By default, Hyper-Threading is enabled. For more information, see Change CPU options.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports the NVMe protocol. For more information, see NVMe protocol.

    • Supports ESSDs, ESSD AutoPL disks, and Regional ESSDs. For information about disks, see Overview of Block Storage.

    • Allows up to 64 data disks to be attached per instance. You can attach up to 16 data disks to an instance when you create the instance. If the instance requires additional data disks, attach more data disks after the instance is created. For more information, see Attach a data disk.

    • Delivers a sequential read/write throughput of up to 64 Gbit/s and up to 1,000,000 IOPS per instance.

    • Provides high storage I/O performance based on large computing capacity. For more information, see Storage I/O performance.

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Provides ultra-high packet forwarding rates.

    • Provides high network performance based on large computing capacity.

r7se instance types

Instance type

vCPUs

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Maximum attached data disks

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.r7se.large

2

16

1.2/burstable up to 3

450,000

Up to 250,000

2

3

6

6

16

30,000/burstable up to 150,000

3/10

ecs.r7se.xlarge

4

32

2/burstable up to 5

500,000

Up to 250,000

4

4

15

15

16

60,000/burstable up to 150,000

4/10

ecs.r7se.2xlarge

8

64

3/burstable up to 8

800,000

Up to 250,000

8

4

15

15

16

100,000/burstable up to 150,000

6/10

ecs.r7se.3xlarge

12

96

4.5/burstable up to 10

1,200,000

Up to 250,000

8

8

15

15

16

120,000/burstable up to 150,000

8/10

ecs.r7se.4xlarge

16

128

6/burstable up to 10

1,500,000

300,000

8

8

30

30

24

150,000/none

10/none

ecs.r7se.6xlarge

24

192

8/burstable up to 10

2,250,000

450,000

12

8

30

30

24

200,000/none

12/none

ecs.r7se.8xlarge

32

256

10/none

3,000,000

600,000

16

8

30

30

30

300,000/none

16/none

ecs.r7se.16xlarge

64

512

16/none

6,000,000

1,200,000

32

8

30

30

56

500,000/none

32/none

g7nex, network-enhanced general-purpose instance family

  • Introduction: This instance family uses the fourth-generation SHENLONG architecture to provide predictable and consistent ultra-high performance. This instance family utilizes fast path acceleration on chips to improve storage performance, network performance, and computing stability by an order of magnitude.

  • Supported scenarios:

    • Network-intensive scenarios such as Network Functions Virtualization (NFV) or Software-defined Wide Area Network (SD-WAN), mobile Internet, live commenting on videos, and telecom data forwarding

    • Small and medium-sized database systems, caches, and search clusters

    • Enterprise-level applications of various types and sizes

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4.

    • Uses the third-generation Intel® Xeon® Scalable (Ice Lake) processors that deliver a base frequency of 2.7 GHz and an all-core turbo frequency of 3.5 GHz to provide consistent computing performance.

    • Supports Hyper-Threading. By default, Hyper-Threading is enabled. For more information, see Change CPU options.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Significantly improves the network throughput and packet forwarding rate per instance. A single instance can deliver a packet forwarding rate of up to 30,000,000 pps.

    • Provides high network performance based on large computing capacity.

g7nex instance types

Instance type

vCPUs

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

EBS queues

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.g7nex.large

2

8

3/burstable up to 20

450,000

2

3

10

10

1

10,000/burstable up to 50,000

1.5/burstable up to 8

ecs.g7nex.xlarge

4

16

5/burstable up to 24

900,000

4

4

15

15

1

20,000/burstable up to 50,000

2/burstable up to 8

ecs.g7nex.2xlarge

8

32

10/burstable up to 32

1,750,000

8

6

15

15

2

25,000/burstable up to 50,000

3/burstable up to 8

ecs.g7nex.4xlarge

16

64

20/burstable up to 40

3,000,000

16

8

30

30

2

40,000/burstable up to 50,000

5/burstable up to 8

ecs.g7nex.8xlarge

32

128

40/none

6,000,000

32

8

30

30

4

75,000/none

8/none

ecs.g7nex.16xlarge

64

256

80/none

8,000,000

32

15

50

50

4

150,000/none

16/none

ecs.g7nex.32xlarge

128

512

160/none

16,000,000

32

15

50

50

4

300,000/none

32/none

Note

Each ecs.g7nex.32xlarge instance must have at least two elastic network interfaces (ENIs) that are assigned different network card indexes before the instance can burst its network bandwidth to 160 Gbit/s. If all ENIs on the instance are assigned the same network card index, the instance can burst its network bandwidth only to 100 Gbit/s. For more information, see AttachNetworkInterface.

c7nex, network-enhanced compute-optimized instance family

  • Introduction: This instance family uses the fourth-generation SHENLONG architecture to provide predictable and consistent ultra-high performance. This instance family utilizes fast path acceleration on chips to improve storage performance, network performance, and computing stability by an order of magnitude.

  • Supported scenarios:

    • Network-intensive scenarios such as NFV or SD-WAN, mobile Internet, live commenting on videos, and telecom data forwarding

    • Small and medium-sized database systems, caches, and search clusters

    • Enterprise-level applications of various types and sizes

  • Compute:

    • Offers a CPU-to-memory ratio of 1:2.

    • Uses the third-generation Intel® Xeon® Scalable (Ice Lake) processors that deliver a base frequency of 2.7 GHz and an all-core turbo frequency of 3.5 GHz to provide consistent computing performance.

    • Supports Hyper-Threading. By default, Hyper-Threading is enabled. For more information, see Change CPU options.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports ESSDs, ESSD AutoPL disks, and Regional ESSDs. For information about disks, see Overview of Block Storage.

    • For small instance types, disk IOPS and disk bandwidth are burstable. The storage I/O performance of an instance corresponds to its instance type. A larger instance type provides higher storage I/O performance. For more information, see Storage I/O performance.

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Significantly improves the network throughput and packet forwarding rate per instance. A single instance can deliver a packet forwarding rate of up to 30,000,000 pps.

    • Provides high network performance based on large computing capacity.

c7nex instance types

Instance type

vCPUs

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

EBS queues

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.c7nex.large

2

4

3/burstable up to 20

450,000

2

3

10

10

1

10,000/burstable up to 50,000

1.5/burstable up to 8

ecs.c7nex.xlarge

4

8

5/burstable up to 24

900,000

4

4

15

15

1

20,000/burstable up to 50,000

2/burstable up to 8

ecs.c7nex.2xlarge

8

16

10/burstable up to 32

1,750,000

8

6

15

15

2

25,000/burstable up to 50,000

3/burstable up to 8

ecs.c7nex.4xlarge

16

32

20/burstable up to 40

3,000,000

16

8

30

30

2

40,000/burstable up to 50,000

5/burstable up to 8

ecs.c7nex.8xlarge

32

64

40/none

6,000,000

32

8

30

30

4

75,000/none

8/none

ecs.c7nex.16xlarge

64

128

80/none

8,000,000

32

15

50

50

4

150,000/none

16/none

ecs.c7nex.32xlarge

128

256

160/none

16,000,000

32

15

50

50

4

300,000/none

32/none

Note

Each ecs.c7nex.32xlarge instance must have at least two ENIs that are assigned different network card indexes before the instance can burst its network bandwidth to 160 Gbit/s. If all ENIs on the instance are assigned the same network card index, the instance can burst its network bandwidth only to 100 Gbit/s. For more information, see AttachNetworkInterface.

g7ne, network-enhanced general-purpose instance family

  • Introduction: This instance family significantly improves the network throughput and packet forwarding rate per instance. A single instance can deliver a packet forwarding rate of up to 24,000,000 pps.

  • Supported scenarios:

    • Network-intensive scenarios such as NFV or SD-WAN, mobile Internet, live commenting on videos, and telecom data forwarding

    • Small and medium-sized database systems, caches, and search clusters

    • Enterprise-level applications of various types and sizes

    • Big data analytics and machine learning

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4.

    • Uses Intel® Xeon® Platinum 8369HB (Cooper Lake) or Intel® Xeon® Platinum 8369HC (Cooper Lake) processors that deliver a turbo frequency of 3.8 GHz and a clock speed of at least 3.3 GHz to provide consistent computing performance.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Provides high network performance based on large computing capacity.

g7ne instance types

Instance type

vCPUs

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.g7ne.large

2

8

1.5/10

900,000

450,000

2

3

10

10

10,000

0.75

ecs.g7ne.xlarge

4

16

3/10

1,000,000

900,000

4

4

15

15

20,000

1

ecs.g7ne.2xlarge

8

32

6/15

1,600,000

1,750,000

8

6

15

15

25,000

1.2

ecs.g7ne.4xlarge

16

64

12/25

3,000,000

3,500,000

16

8

30

30

40,000

2

ecs.g7ne.8xlarge

32

128

25/none

6,000,000

6,000,000

32

8

30

30

75,000

5

ecs.g7ne.12xlarge

48

192

40/none

12,000,000

8,000,000

32

8

30

30

100,000

8

ecs.g7ne.24xlarge

96

384

80/none

24,000,000

16,000,000

48

15

50

50

240,000

16

g5ne, network-enhanced general-purpose instance family

  • Introduction: This instance family significantly improves the network throughput and packet forwarding rate per instance.

  • Supported scenarios:

    • Data Plane Development Kit (DPDK) applications

    • Network-intensive scenarios such as NFV or SD-WAN, mobile Internet, live commenting on videos, and telecom data forwarding

    • Small and medium-sized database systems, caches, and search clusters

    • Enterprise-level applications of various types and sizes

    • Big data analytics and machine learning

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4.

    • Uses 2.5 GHz Intel® Xeon® Platinum 8163 (Skylake) or 8269CY (Cascade Lake) processors to provide consistent computing performance.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports standard SSDs and ultra disks. For information about disks, see Overview of Block Storage.

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Provides high network performance based on large computing capacity.

      Note

      To deploy DPDK applications, we recommend that you select instance types in the g5ne instance family.

g5ne instance types

Instance type

vCPUs

Memory (GiB)

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.g5ne.large

2

8

1

400,000

450,000

2

3

10

10

10,000

1

ecs.g5ne.xlarge

4

16

2

750,000

900,000

4

4

15

15

15,000

1

ecs.g5ne.2xlarge

8

32

3.5

1,500,000

1,750,000

8

6

15

15

30,000

1

ecs.g5ne.4xlarge

16

64

7

3,000,000

3,500,000

16

8

30

30

60,000

2

ecs.g5ne.8xlarge

32

128

15

6,000,000

7,000,000

32

8

30

30

110,000

4

ecs.g5ne.16xlarge

64

256

30

12,000,000

14,000,000

32

8

30

30

130,000

8

ecs.g5ne.18xlarge

72

288

33

13,500,000

15,000,000

32

15

50

50

160,000

9

g9it, security-enhanced general-purpose instance family

  • Introduction:

    • Supports Intel® SGX encrypted computing with up to 192 GiB of encrypted memory. This ensures the confidentiality and integrity of critical code and data against malware attacks.

    • Supports SGX technology in virtual machine form factors. You can choose instance types as needed.

    • Hyper-threading is disabled by default, providing exclusive physical cores. This reduces the risk of side-channel attacks. The memory encryption algorithm is upgraded to AES-256.

      Important

      If you use keys (such as SGX sealing keys) that are bound to hardware to encrypt the data of an instance within an Intel SGX enclave, the encrypted data cannot be decrypted after the host of the instance is changed. We recommend that you perform data redundancy and backup at the application layer to ensure application reliability.

    • Leveraging TPM/TCM chips, the boot chain, from the underlying server hardware to the GuestOS, is measured and authenticated. This enables trusted boot.

    • Uses the new Alibaba Cloud CIPU architecture, offloading many virtualization functions to dedicated hardware. This reduces virtualization overhead and delivers stable, predictable, ultra-high performance.

  • Scenarios:

    • Scenarios involving sensitive information, such as personally identifiable information (PII), healthcare data, financial data, and intellectual property data.

    • Multi-party computation requiring shared confidential data.

    • Blockchain scenarios.

    • Confidential machine learning.

    • Scenarios with high security and trust requirements, such as finance, government, and enterprise applications.

    • Enterprise applications of all types and scales.

  • Compute:

    • Processor-to-memory ratio is 1:4. Encrypted memory accounts for approximately 50% of the total memory.

    • Processor: Intel® Xeon® Granite Rapids, with a clock speed of 3.2 GHz and an all-core turbo frequency of 3.6 GHz. This provides stable compute performance.

  • Storage:

  • Networking:

    • Supports IPv4 and IPv6. For more information, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Instance network performance scales with instance specifications. Higher specifications provide stronger network performance.

g9it includes the instance types and metric data listed in the following table.

Instance type

vCPUs

Memory (GiB)

Encrypted memory (GiB)

Network bandwidth baseline/burst (Gbit/s)

Packet forwarding PPS

Supports vTPM

Connections

NIC queues

ENIs

Private IPv4 addresses per NIC

IPv6 addresses per NIC

Disk baseline/burst IOPS

Disk bandwidth baseline/burst (Gbit/s)

ecs.g9it.large

2

8

4

4/15

1,200,000

Yes

500,000

4

4

15

15

50,000/200,000

2.5/10

ecs.g9it.xlarge

4

16

8

6/15

1,600,000

Yes

500,000

8

4

15

15

60,000/200,000

4/10

ecs.g9it.2xlarge

8

32

16

12/25

3,000,000

Yes

500,000

16

8

30

30

100,000/200,000

6/10

ecs.g9it.4xlarge

16

64

32

20/32

6,000,000

Yes

800,000

32

8

30

30

200,000/300,000

10/12

ecs.g9it.8xlarge

32

128

64

28/36

12,000,000

Yes

2,000,000

64

8

30

30

300,000/400,000

16/24

ecs.g9it.16xlarge

64

256

128

36/50

20,000,000

Yes

4,000,000

64

15

30

30

400,000/650,000

24/28

ecs.g9it.24xlarge

96

384

192

64

24,000,000

Yes

6,000,000

64

15

50

50

500,000/800,000

32

Note
  • Intel® Xeon® Granite Rapids supports only Intel SGX DCAP–based remote attestation and does not support Intel EPID–based remote attestation. You may need to adapt your program to use the remote attestation feature normally. For more information about remote attestation, see attestation-service.

  • Intel SGX depends on host hardware. This instance family does not support hot migration.

  • Operations, such as changing instance types and enabling the economical mode, may cause the host of an instance to change. For instances of this instance family, the host change may cause data decryption to fail. Proceed with caution.

  • By default, failover is disabled. You can enable failover. For more information, see Modify maintenance attributes. Failover causes the host of an instance to change. For instances of this instance family, the host change may cause data decryption to fail. Proceed with caution.

  • When you create a security-enhanced instance, you must select a dedicated image to use the security features. For more information, see Create trusted instance.

  • The product is in invitational preview. Submit a ticket to request access.

r9it, security-enhanced memory-optimized instance family

  • Introduction

    • Supports Intel® SGX encrypted computing with up to 384 GiB of encrypted memory. This protects the confidentiality and integrity of critical code and data from malware attacks.

    • Supports SGX technology in virtual machine form. You can choose instance types flexibly based on your needs.

    • Hyper-threading is disabled by default. Each instance is allocated dedicated physical CPU cores, reducing side-channel attack risks. The memory encryption algorithm is upgraded to AES-256.

      Important

      If you use keys (such as SGX sealing keys) that are bound to hardware to encrypt the data of an instance within an Intel SGX enclave, the encrypted data cannot be decrypted after the host of the instance is changed. We recommend that you perform data redundancy and backup at the application layer to ensure application reliability.

    • Uses TPM/TCM chips to measure and validate the boot chain—from underlying server hardware to GuestOS—enabling trusted boot.

    • Uses Alibaba Cloud’s new Cloud Infrastructure Processing Unit (CIPU) architecture. This offloads many virtualization functions to dedicated hardware, reducing virtualization overhead and delivering stable, predictable high performance.

  • Scenarios

    • Workloads handling sensitive information such as personally identifiable information, healthcare data, financial data, and intellectual property.

    • Secure multi-party computation where confidential data must be shared.

    • Blockchain applications.

    • Confidential machine learning.

    • High-security, high-trust environments such as finance, government, and enterprise systems.

    • Enterprise-grade applications of all types and scales.

  • Compute

    • Processor-to-memory ratio is 1:8. Encrypted memory accounts for about 50% of total memory.

    • Processor: Intel® Xeon® Granite Rapids, base clock speed 3.2 GHz, all-core turbo frequency 3.6 GHz. Delivers stable compute performance.

  • Storage

  • Network

    • Supports IPv4 and IPv6. For IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Network performance scales with instance type. Higher-spec instances deliver stronger network performance.

r9it includes the instance types and metric data listed in the following table.

Instance type

vCPUs

Memory (GiB)

Encrypted memory (GiB)

Baseline/burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

vTPM support

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.r9it.large

2

16

8

4/15

1,200,000

Yes

500,000

4

4

15

15

50,000/200,000

2.5/10

ecs.r9it.xlarge

4

32

16

6/15

1,600,000

Yes

500,000

8

4

15

15

60,000/200,000

4/10

ecs.r9it.2xlarge

8

64

32

12/25

3,000,000

Yes

500,000

16

8

30

30

100,000/200,000

6/10

ecs.r9it.4xlarge

16

128

64

20/32

6,000,000

Yes

800,000

32

8

30

30

200,000/300,000

10/12

ecs.r9it.8xlarge

32

256

128

28/36

12,000,000

Yes

2,000,000

64

8

30

30

300,000/400,000

16/24

ecs.r9it.16xlarge

64

512

256

36/50

20,000,000

Yes

4,000,000

64

15

30

30

400,000/650,000

24/28

ecs.r9it.24xlarge

96

768

384

64

24,000,000

Yes

6,000,000

64

15

50

50

500,000/800,000

32

Note
  • Intel® Xeon® Granite Rapids supports only Intel SGX DCAP-based remote attestation—not Intel EPID-based remote attestation. You may need to adapt your application before using remote attestation. For more information, see attestation-service.

  • Intel SGX depends on host hardware. This instance family does not support hot migration.

  • Operations, such as changing instance types and enabling the economical mode, may cause the host of an instance to change. For instances of this instance family, the host change may cause data decryption to fail. Proceed with caution.

  • By default, failover is disabled. You can enable failover. For more information, see Modify maintenance attributes. Failover causes the host of an instance to change. For instances of this instance family, the host change may cause data decryption to fail. Proceed with caution.

  • When you create a security-enhanced instance, you must select a dedicated image to use the security features. For more information, see Create trusted instance.

  • This product is in invitational preview. To request access, or submit a ticket.

g7t, security-enhanced general-purpose instance family

  • Introduction:

    • This instance family supports up to 256 GiB of encrypted memory and confidential computing based on Intel® Software Guard Extensions (SGX) to protect the confidentiality and integrity of essential code and data from malware attacks.

    • This instance family supports Virtual SGX (vSGX) and allows you to select instance types based on your business requirements.

      Important

      If you use keys (such as SGX sealing keys) that are bound to hardware to encrypt the data of an instance within an Intel SGX enclave, the encrypted data cannot be decrypted after the host of the instance is changed. We recommend that you perform data redundancy and backup at the application layer to ensure application reliability.

    • This instance family implements trusted boot based on Trusted Cryptography Module (TCM) or Trusted Platform Module (TPM) chips. During a trusted boot, all modules in the boot chain from the underlying server to the guest operating system are measured and verified.

    • This instance family offloads a large number of virtualization features to dedicated hardware by using the third-generation SHENLONG architecture to provide predictable and consistent ultra-high performance and reduce virtualization overheads.

  • Supported scenarios:

    • Scenarios that involve sensitive information such as personal identity information, healthcare information, financial information, and intellectual property data

    • Scenarios in which confidential data is shared among multiple parties

    • Blockchain scenarios

    • Confidential machine learning

    • Scenarios that require high security and enhanced trust, such as services for financial organizations, public service sectors, and enterprises

    • Enterprise-level applications of various types and sizes

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4. About 50% of memory is encrypted.

    • Uses the third-generation Intel® Xeon® Scalable (Ice Lake) processors that deliver a base frequency of 2.7 GHz and an all-core turbo frequency of 3.5 GHz to provide consistent computing performance.

    • Supports Hyper-Threading. By default, Hyper-Threading is enabled. For more information, see Change CPU options.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances provide an ultra-high packet forwarding rate.

    • Provides high network performance based on large computing capacity.

g7t instance types

Instance type

vCPUs

Memory (GiB)

Encrypted memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Support for vTPM

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.g7t.large

2

8

4

2/burstable up to 10

900,000

Yes

Up to 250,000

2

3

6

6

20,000/burstable up to 110,000

1.5/burstable up to 6

ecs.g7t.xlarge

4

16

8

3/burstable up to 10

1,000,000

Yes

Up to 250,000

4

4

15

15

40,000/burstable up to 110,000

2/burstable up to 6

ecs.g7t.2xlarge

8

32

16

5/burstable up to 10

1,600,000

Yes

Up to 250,000

8

4

15

15

50,000/burstable up to 110,000

3/burstable up to 6

ecs.g7t.3xlarge

12

48

24

8/burstable up to 10

2,400,000

Yes

Up to 250,000

8

8

15

15

70,000/burstable up to 110,000

4/burstable up to 6

ecs.g7t.4xlarge

16

64

32

10/burstable up to 25

3,000,000

Yes

300,000

8

8

30

30

80,000/burstable up to 110,000

5/burstable up to 6

ecs.g7t.6xlarge

24

96

48

12/burstable up to 25

4,500,000

Yes

450,000

12

8

30

30

110,000/none

6/none

ecs.g7t.8xlarge

32

128

64

16/burstable up to 25

6,000,000

Yes

600,000

16

8

30

30

150,000/none

8/none

ecs.g7t.16xlarge

64

256

128

32/none

12,000,000

Yes

1,200,000

32

8

30

30

300,000/none

16/none

ecs.g7t.32xlarge

128

512

256

64/none

24,000,000

Yes

2,400,000

32

15

30

30

600,000/none

32/none

Note
  • Intel Ice Lake supports only remote attestation based on Intel Software Guard Extensions Data Center Attestation Primitives (Intel SGX DCAP) and does not support remote attestation based on Intel Enhanced Privacy ID (EPID). You must adapt applications before you can use the remote attestation feature. For more information about remote attestation, see Strengthen Enclave Trust with Attestation.

  • Intel SGX depends on host hardware. This instance family does not support hot migration.

  • Operations, such as changing instance types and enabling the economical mode, may cause the host of an instance to change. For instances of this instance family, the host change may cause data decryption to fail. Proceed with caution.

  • By default, failover is disabled. You can enable failover. For more information, see Modify maintenance attributes. Failover causes the host of an instance to change. For instances of this instance family, the host change may cause data decryption to fail. Proceed with caution.

  • When you create a security-enhanced instance, you must select a dedicated image to use the security features. For more information, see Create trusted instance.

  • To use the ecs.g7t.32xlarge instance type, submit a ticket.

c7t, security-enhanced compute-optimized instance family

  • Introduction:

    • This instance family supports up to 128 GiB of encrypted memory and confidential computing based on Intel® SGX to protect the confidentiality and integrity of essential code and data from malware attacks.

    • This instance family supports vSGX and allows you to select instance types based on your business requirements.

      Important

      If you use keys (such as SGX sealing keys) that are bound to hardware to encrypt the data of an instance within an Intel SGX enclave, the encrypted data cannot be decrypted after the host of the instance is changed. We recommend that you perform data redundancy and backup at the application layer to ensure application reliability.

    • This instance family implements trusted boot based on TCM or TPM chips. During a trusted boot, all modules in the boot chain from the underlying server to the guest operating system are measured and verified.

    • This instance family offloads a large number of virtualization features to dedicated hardware by using the third-generation SHENLONG architecture to provide predictable and consistent ultra-high performance and reduce virtualization overheads.

  • Supported scenarios:

    • Scenarios that involve sensitive information such as personal identity information, healthcare information, financial information, and intellectual property data

    • Scenarios in which confidential data is shared among multiple parties

    • Blockchain scenarios

    • Confidential machine learning

    • Scenarios that require high security and enhanced trust, such as services for financial organizations, public service sectors, and enterprises

    • Enterprise-level applications of various types and sizes

  • Compute:

    • Offers a CPU-to-memory ratio of 1:2. About 50% of memory is encrypted.

    • Uses the third-generation Intel® Xeon® Scalable (Ice Lake) processors that deliver a base frequency of 2.7 GHz and an all-core turbo frequency of 3.5 GHz to provide consistent computing performance.

    • Supports Hyper-Threading. By default, Hyper-Threading is enabled. For more information, see Change CPU options.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances provide an ultra-high packet forwarding rate.

    • Provides high network performance based on large computing capacity.

c7t instance types

Instance type

vCPUs

Memory (GiB)

Encrypted memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Support for vTPM

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.c7t.large

2

4

2

2/burstable up to 10

900,000

Yes

Up to 250,000

2

3

6

6

20,000/burstable up to 110,000

1.5/burstable up to 6

ecs.c7t.xlarge

4

8

4

3/burstable up to 10

1,000,000

Yes

Up to 250,000

4

4

15

15

40,000/burstable up to 110,000

2/burstable up to 6

ecs.c7t.2xlarge

8

16

8

5/burstable up to 10

1,600,000

Yes

Up to 250,000

8

4

15

15

50,000/burstable up to 110,000

3/burstable up to 6

ecs.c7t.3xlarge

12

24

12

8/burstable up to 10

2,400,000

Yes

Up to 250,000

8

8

15

15

70,000/burstable up to 110,000

4/burstable up to 6

ecs.c7t.4xlarge

16

32

16

10/burstable up to 25

3,000,000

Yes

300,000

8

8

30

30

80,000/burstable up to 110,000

5/burstable up to 6

ecs.c7t.6xlarge

24

48

24

12/burstable up to 25

4,500,000

Yes

450,000

12

8

30

30

110,000/none

6/none

ecs.c7t.8xlarge

32

64

32

16/burstable up to 25

6,000,000

Yes

600,000

16

8

30

30

150,000/none

8/none

ecs.c7t.16xlarge

64

128

64

32/none

12,000,000

Yes

1,200,000

32

8

30

30

300,000/none

16/none

ecs.c7t.32xlarge

128

256

128

64/none

24,000,000

Yes

2,400,000

32

15

30

30

600,000/none

32/none

Note
  • Intel Ice Lake supports only remote attestation based on Intel Software Guard Extensions Data Center Attestation Primitives (Intel SGX DCAP) and does not support remote attestation based on Intel Enhanced Privacy ID (EPID). You must adapt applications before you can use the remote attestation feature. For more information about remote attestation, see Strengthen Enclave Trust with Attestation.

  • Intel SGX depends on host hardware. This instance family does not support hot migration.

  • Operations, such as changing instance types and enabling the economical mode, may cause the host of an instance to change. For instances of this instance family, the host change may cause data decryption to fail. Proceed with caution.

  • By default, failover is disabled. You can enable failover. For more information, see Modify maintenance attributes. Failover causes the host of an instance to change. For instances of this instance family, the host change may cause data decryption to fail. Proceed with caution.

  • When you create a security-enhanced instance, you must select a dedicated image to use the security features. For more information, see Create trusted instance.

  • To use the ecs.c7t.32xlarge instance type, submit a ticket.

r7t, security-enhanced memory-optimized instance family

  • Introduction:

    • This instance family supports up to 512 GiB of encrypted memory and confidential computing based on Intel® SGX to protect the confidentiality and integrity of essential code and data from malware attacks.

    • This instance family supports vSGX and allows you to select instance types based on your business requirements.

      Important

      If you use keys (such as SGX sealing keys) that are bound to hardware to encrypt the data of an instance within an Intel SGX enclave, the encrypted data cannot be decrypted after the host of the instance is changed. We recommend that you perform data redundancy and backup at the application layer to ensure application reliability.

    • This instance family implements trusted boot based on TCM or TPM chips. During a trusted boot, all modules in the boot chain from the underlying server to the guest operating system are measured and verified.

    • This instance family offloads a large number of virtualization features to dedicated hardware by using the third-generation SHENLONG architecture to provide predictable and consistent ultra-high performance and reduce virtualization overheads.

  • Supported scenarios:

    • Encrypted computing applications for databases

    • Scenarios that involve sensitive information such as personal identity information, healthcare information, financial information, and intellectual property data

    • Scenarios in which confidential data is shared among multiple parties

    • Blockchain scenarios

    • Confidential machine learning

    • Scenarios that require high security and enhanced trust, such as services for financial organizations, public service sectors, and enterprises

    • Enterprise-level applications of various types and sizes

  • Compute:

    • Offers a CPU-to-memory ratio of 1:8. About 50% of memory is encrypted.

    • Uses the third-generation Intel® Xeon® Scalable (Ice Lake) processors that deliver a base frequency of 2.7 GHz and an all-core turbo frequency of 3.5 GHz to provide consistent computing performance.

    • Supports Hyper-Threading. By default, Hyper-Threading is enabled. For more information, see Change CPU options.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Provides ultra-high packet forwarding rates.

    • Provides high network performance based on large computing capacity.

r7t instance types

Instance type

vCPUs

Memory (GiB)

Encrypted memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Support for vTPM

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.r7t.large

2

16

8

2/burstable up to 10

900,000

Yes

Up to 250,000

2

3

6

6

20,000/burstable up to 110,000

1.5/burstable up to 6

ecs.r7t.xlarge

4

32

16

3/burstable up to 10

1,000,000

Yes

Up to 250,000

4

4

15

15

40,000/burstable up to 110,000

2/burstable up to 6

ecs.r7t.2xlarge

8

64

32

5/burstable up to 10

1,600,000

Yes

Up to 250,000

8

4

15

15

50,000/burstable up to 110,000

3/burstable up to 6

ecs.r7t.3xlarge

12

96

48

8/burstable up to 10

2,400,000

Yes

Up to 250,000

8

8

15

15

70,000/burstable up to 110,000

4/burstable up to 6

ecs.r7t.4xlarge

16

128

64

10/burstable up to 25

3,000,000

Yes

300,000

8

8

30

30

80,000/burstable up to 110,000

5/burstable up to 6

ecs.r7t.6xlarge

24

192

96

12/burstable up to 25

4,500,000

Yes

450,000

12

8

30

30

110,000/none

6/none

ecs.r7t.8xlarge

32

256

128

16/burstable up to 25

6,000,000

Yes

600,000

16

8

30

30

150,000/none

8/none

ecs.r7t.16xlarge

64

512

256

32/none

12,000,000

Yes

1,200,000

32

8

30

30

300,000/none

16/none

ecs.r7t.32xlarge

128

1,024

512

64/none

24,000,000

Yes

2,400,000

32

15

30

30

600,000/none

32/none

Note
  • Intel Ice Lake supports only remote attestation based on Intel Software Guard Extensions Data Center Attestation Primitives (Intel SGX DCAP) and does not support remote attestation based on Intel Enhanced Privacy ID (EPID). You must adapt applications before you can use the remote attestation feature. For more information about remote attestation, see Strengthen Enclave Trust with Attestation.

  • Intel SGX depends on host hardware. This instance family does not support hot migration.

  • Operations, such as changing instance types and enabling the economical mode, may cause the host of an instance to change. For instances of this instance family, the host change may cause data decryption to fail. Proceed with caution.

  • By default, failover is disabled. You can enable failover. For more information, see Modify maintenance attributes. Failover causes the host of an instance to change. For instances of this instance family, the host change may cause data decryption to fail. Proceed with caution.

  • When you create a security-enhanced instance, you must select a dedicated image to use the security features. For more information, see Create trusted instance.

  • To use the ecs.r7t.32xlarge instance type, submit a ticket.

g6t, security-enhanced general-purpose instance family

Features:

  • Introduction:

    • This instance family implements trusted boot based on TCM or TPM chips. During a trusted boot, all modules in the boot chain from the underlying server to the guest operating system are measured and verified.

    • This instance family supports the vTPM feature and delivers trusted capabilities at the IaaS layer based on integrity monitoring.

    • This instance family offloads a large number of virtualization features to dedicated hardware by using the third-generation SHENLONG architecture to provide predictable and consistent ultra-high performance and reduce virtualization overheads. This instance family utilizes fast path acceleration on chips to improve storage performance, network performance, and computing stability by an order of magnitude.

  • Supported scenarios:

    • Scenarios that require high security and enhanced trust, such as services for financial organizations, public service sectors, and enterprises

    • Scenarios where large volumes of packets are received and transmitted, such as live commenting on videos and telecom data forwarding

    • Enterprise-level applications of various types and sizes

    • Websites and application servers

    • Game servers

    • Small and medium-sized database systems, caches, and search clusters

    • Data analytics and computing

    • Computing clusters and memory-intensive data processing

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4.

    • Uses 2.5 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processors that deliver a turbo frequency of 3.2 GHz to provide consistent computing performance.

    • Supports Hyper-Threading. By default, Hyper-Threading is enabled. For more information, see Change CPU options.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Provides ultra-high packet forwarding rates.

    • Provides high network performance based on large computing capacity.

g6t instance types

Instance type

vCPUs

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Support for vTPM

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.g6t.large

2

8

1.2/burstable up to 10

900,000

Yes

Up to 250,000

2

3

6

1

20,000

1

ecs.g6t.xlarge

4

16

2/burstable up to 10

1,000,000

Yes

Up to 250,000

4

4

15

1

40,000

1.5

ecs.g6t.2xlarge

8

32

3/burstable up to 10

1,600,000

Yes

Up to 250,000

8

4

15

1

50,000

2

ecs.g6t.4xlarge

16

64

6/burstable up to 10

3,000,000

Yes

300,000

8

8

30

1

80,000

3

ecs.g6t.8xlarge

32

128

10/none

6,000,000

Yes

600,000

16

8

30

1

150,000

5

ecs.g6t.13xlarge

52

192

16/none

9,000,000

Yes

900,000

32

7

30

1

240,000

8

ecs.g6t.26xlarge

104

384

32/none

24,000,000

Yes

1,800,000

32

15

30

1

480,000

16

Note

The results for network capabilities are the maximum values obtained from single-item tests. For example, when network bandwidth is tested, no stress tests are performed on the packet forwarding rate or other network metrics.

c6t, security-enhanced compute-optimized instance family

  • Introduction:

    • This instance family implements trusted boots based on TPM chips. During a trusted boot, all modules in the boot chain from the underlying hardware to the guest operating system are measured and verified.

    • This instance family supports integrity monitoring and provides trusted capabilities at the IaaS layer.

    • This instance family offloads a large number of virtualization features to dedicated hardware by using the third-generation SHENLONG architecture to provide predictable and consistent ultra-high performance and reduce virtualization overheads. This instance family utilizes fast path acceleration on chips to improve storage performance, network performance, and computing stability by an order of magnitude.

  • Supported scenarios:

    • Scenarios that require high security and enhanced trust, such as services for financial organizations, public service sectors, and enterprises

    • Scenarios where large volumes of packets are received and transmitted, such as live commenting on videos and telecom data forwarding

    • Web frontend servers

    • Frontend servers of massively multiplayer online (MMO) games

    • Data analytics, batch processing, and video encoding

    • High-performance scientific and engineering applications

  • Compute:

    • Offers a CPU-to-memory ratio of 1:2.

    • Uses 2.5 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processors that deliver a turbo frequency of 3.2 GHz to provide consistent computing performance.

    • Supports Hyper-Threading. By default, Hyper-Threading is enabled. For more information, see Change CPU options.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Provides ultra-high packet forwarding rates.

    • Provides high network performance based on large computing capacity.

c6t instance types

Instance type

vCPUs

Memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Support for vTPM

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.c6t.large

2

4

1.2/burstable up to 10

900,000

Yes

Up to 250,000

2

3

6

1

20,000

1

ecs.c6t.xlarge

4

8

2/burstable up to 10

1,000,000

Yes

Up to 250,000

4

4

15

1

40,000

1.5

ecs.c6t.2xlarge

8

16

3/burstable up to 10

1,600,000

Yes

Up to 250,000

8

4

15

1

50,000

2

ecs.c6t.4xlarge

16

32

6/burstable up to 10

3,000,000

Yes

300,000

8

8

30

1

80,000

3

ecs.c6t.8xlarge

32

64

10/none

6,000,000

Yes

600,000

16

8

30

1

150,000

5

ecs.c6t.13xlarge

52

96

16/none

9,000,000

Yes

900,000

32

7

30

1

240,000

8

ecs.c6t.26xlarge

104

192

32/none

24,000,000

Yes

1,800,000

32

15

30

1

480,000

16

Note

The results for network capabilities are the maximum values obtained from single-item tests. For example, when network bandwidth is tested, no stress tests are performed on the packet forwarding rate or other network metrics.

re8, high-memory instance family

  • Introduction: This family uses the new Alibaba Cloud Cloud Infrastructure Processing Unit (CIPU) architecture to deliver stable and predictable ultra-high performance. It also uses chip-level fast path acceleration to significantly improve storage performance, network performance, and computing stability.

  • Supported scenarios: In-memory databases such as SAP HANA, high-performance databases, and other memory-intensive enterprise applications.

  • Compute:

    • Processor-to-memory ratio of 1:17. Maximum memory capacity of 16 TB.

    • Processor: Intel ® Xeon ® Sapphire Rapids processors with a base clock speed of 1.9 GHz and an all-core turbo frequency of 2.9 GHz. Provides stable compute performance.

  • Storage:

    • This is an I/O optimized instance.

    • Supports the NVMe protocol. For more information, see NVMe protocol.

    • Supported disk types: ESSDs , ESSD AutoPL disk , and Regional ESSDs. For more information about disks, see Overview of Block Storage.

    • Smaller instance types support burst disk IOPS and bandwidth. The storage I/O performance of an instance depends on its instance type. Instances with higher specifications deliver better storage I/O performance. For more information, see Storage I/O performance.

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • Supports Jumbo Frames. For more information, see Jumbo Frames.

    • Instance network performance scales with instance type. Higher instance types deliver stronger network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

re8 instance types

Instance type

vCPUs

Memory (GiB)

Base/Burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

NIC Queues

Elastic Network Interfaces (ENIs)

Private IPv4 addresses per ENI

Base disk IOPS

Base disk bandwidth (Gbit/s)

ecs.re8.30xlarge

120

2,048

32/Up to 48

7,500,000

32

15

30

150,000/Up to 300,000

12/Up to 25

ecs.re8.60xlarge

240

4,096

64

15,000,000

32

15

30

300,000

25

ecs.re8.90xlarge

360

6,144

96

22,500,000

64

15

30

400,000

36

ecs.re8.120xlarge

480

8,192

128

30,000,000

64

15

40

600,000

50

ecs.re8.180xlarge

720

12,288

192

45,000,000

64

15

40

900,000

75

ecs.re8.240xlarge

960

16,384

200

50,000,000

64

15

50

1,200,000

100

re6p, persistent memory-optimized instance family

For answers to commonly asked questions about persistent memory-optimized instances, see Instance FAQ.

Features:

  • Introduction:

    • This instance family uses Intel® OptaneTM persistent memory.

      Important

      The reliability of data stored in persistent memory varies based on the reliability of persistent memory devices and the physical servers to which these devices are attached. Risks of single points of failure exist. To ensure the reliability of application data, we recommend that you implement data redundancy at the application layer and use cloud disks for long-term data storage.

    • This instance family allows persistent memory to be used as memory or as local SSDs on instances of some instance types.

      Note

      For more information, see Configure the usage mode of persistent memory.

    • This instance family provides the ecs.re6p-redis.<nx>large instance types for Redis applications.

      Note

      ecs.re6p-redis.<nx>large instance types are exclusively provided for Redis applications. Persistent memory on instances of these instance types is used as memory by default and cannot be re-configured as local SSDs. For information about how to deploy a Redis application, see Deploy Redis applications on instances with persistent memory.

  • Supported scenarios:

    • Redis and other NoSQL databases such as Cassandra and MongoDB

    • Structured databases such as MySQL

    • I/O-intensive applications such as e-commerce, online games, and media applications

    • Search scenarios that use solutions such as Elasticsearch

    • Live video streaming, instant messaging, and room-based online games that require persistent connections

    • High-performance relational databases and OLTP systems

  • Compute:

    • Uses 2.5 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processors that deliver a turbo frequency of 3.2 GHz to provide consistent computing performance.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

re6p instance types

Instance type

vCPUs

Memory (GiB)

Persistent memory (GiB)

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

Connections

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.re6p.large

2

8

31.5

1/3

300,000

Up to 250,000

2

2

6

1

10,000

1

ecs.re6p.xlarge

4

16

63

1.5/5

500,000

Up to 250,000

4

3

10

1

20,000

1.5

ecs.re6p.2xlarge

8

32

126

2.5/none

800,000

Up to 250,000

8

4

20

1

25,000

2

ecs.re6p.13xlarge

52

192

756

12.5/none

3,000,000

900,000

32

7

20

1

100,000

8

ecs.re6p.26xlarge

104

384

1,512

25/none

6,000,000

1,800,000

32

15

20

1

200,000

16.4

ecs.re6p-redis.large

2

8

31.5

1/3

300,000

Up to 250,000

2

2

6

1

10,000

1

ecs.re6p-redis.xlarge

4

16

63

1.5/5

500,000

Up to 250,000

4

3

10

1

20,000

1.5

ecs.re6p-redis.2xlarge

8

32

126

2.5/none

800,000

Up to 250,000

8

4

20

1

25,000

2

ecs.re6p-redis.13xlarge

52

192

756

12.5/none

3,000,000

900,000

32

7

20

1

100,000

8

re6, high-memory instance family

Features:

  • Introduction: This instance family is optimized for high-performance databases, in-memory databases, and enterprise-level memory-intensive applications.

  • Supported scenarios:

    • High-performance databases and in-memory databases such as SAP HANA

    • Memory-intensive applications

    • Big data processing engines such as Apache Spark and Presto

  • Compute:

    • Offers a CPU-to-memory ratio of 1:16 (1:15 for some instance types) and up to 3 TiB of memory.

    • Uses 2.5 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processors that deliver a turbo frequency of 3.2 GHz to provide consistent computing performance.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

re6 instance types

Instance type

vCPUs

Memory (GiB)

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (pps)

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.re6.4xlarge

16

256

5

1,800,000

8

7

20

1

25,000

2

ecs.re6.8xlarge

32

512

10

1,800,000

16

7

20

1

50,000

4

ecs.re6.13xlarge

52

768

10

1,800,000

16

7

20

1

50,000

4

ecs.re6.16xlarge

64

1,024

16

3,000,000

32

7

20

1

100,000

8

ecs.re6.26xlarge

104

1,536

16

3,000,000

32

7

20

1

100,000

8

ecs.re6.32xlarge

128

2,048

32

6,000,000

32

15

20

1

200,000

16

ecs.re6.52xlarge

208

3,072

32

6,000,000

32

15

20

1

200,000

16

Note

To use the ecs.re6.32xlarge instance type, submit a ticket.

re4, high-memory instance family

  • Introduction:

    • This instance family is optimized for high-performance databases, in-memory databases, and enterprise-level memory-intensive applications.

    • The ecs.re4.20xlarge and ecs.re4.40xlarge instance types are SAP HANA-certified.

  • Supported scenarios:

    • High-performance databases and in-memory databases such as SAP HANA

    • Memory-intensive applications

    • Big data processing engines such as Apache Spark and Presto

  • Compute:

    • Offers a CPU-to-memory ratio of 1:12 and up to 1,920 GiB of memory.

    • Uses 2.2 GHz Intel® Xeon® E7 8880 v4 (Broadwell) processors that deliver a turbo frequency of up to 2.4 GHz to provide consistent computing performance.

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports standard SSDs and ultra disks. For information about disks, see Overview of Block Storage.

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

re4 instance types

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Packet forwarding rate (pps)

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

ecs.re4.10xlarge

40

480

8

1,000,000

8

4

10

1

ecs.re4.20xlarge

80

960

15

2,000,000

16

2

10

1

ecs.re4.40xlarge

160

1,920

30

4,000,000

16

2

10

1

re4e, high-memory instance family

To use the re4e instance family, submit a ticket.

  • Introduction: This instance family is optimized for high-performance databases, in-memory databases, and enterprise-level memory-intensive applications.

  • Compute:

    • Offers a CPU-to-memory ratio of 1:24 and up to 3,840 GiB of memory.

    • Uses 2.2 GHz Intel® Xeon® E7 8880 v4 (Broadwell) processors that deliver a turbo frequency of up to 2.4 GHz to provide consistent computing performance.

  • Supported scenarios:

    • High-performance databases and in-memory databases such as SAP HANA

    • Memory-intensive applications

    • Big data processing engines such as Apache Spark and Presto

  • Storage:

    • Is an instance family in which all instances are I/O optimized.

    • Supports standard SSDs and ultra disks. For information about disks, see Overview of Block Storage.

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

re4e instance types

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Packet forwarding rate (pps)

NIC queues

ENIs

Private IPv4 addresses per ENI

Private IPv6 addresses per ENI

ecs.re4e.40xlarge

160

3,840

30

4,500,000

16

15

10

1

Entry-level x86 instance families

e, economy instance family

  • Scenarios: Small and medium-sized websites, development and testing environments, and lightweight applications.

  • Compute:

    • Supports multiple processor-to-memory ratios, including 1:1, 1:2, and 1:4.

    • Processor: Intel® Xeon® Platinum scalable processor.

      Note

      e instances use a non-dedicated CPU scheduling mode where each vCPU is randomly assigned to an available CPU hyper-thread. Compared to enterprise-level instances, e instances prioritize resource sharing and are more cost-effective.

  • Storage:

    • I/O optimized instance.

    • Can be attached only with ESSD Entry disks (recommended), enterprise SSDs (ESSDs), and ESSD AutoPL disks.

      Note

      Due to the limitations of economy instance types, ESSDs at performance level (PL) 1, PL2, and PL3 cannot deliver their maximum performance. We recommend that you use ESSD Entry disks or PL0 ESSDs.

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Can be deployed only in virtual private clouds (VPCs).

    • The network performance of an instance is proportional to its compute specifications. Larger instance types provide higher network performance.

The following table lists the available instance types and their performance metrics:

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Multi-Queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Disk baseline/burst IOPS

Disk baseline/burst bandwidth (Gbit/s)

ecs.e-c4m1.large

2

0.5

0.2/Up to 2

1

2

2

1

8,000/N/A

0.4/N/A

ecs.e-c2m1.large

2

1

0.2/Up to 2

1

2

2

1

8,000/N/A

0.4/N/A

ecs.e-c1m1.large

2

2.0

0.2/Up to 2

1

2

2

1

8,000/N/A

0.4/N/A

ecs.e-c1m2.large

2

4.0

0.2/Up to 2

1

2

2

1

8,000/N/A

0.4/N/A

ecs.e-c1m4.large

2

8.0

0.4/Up to 2

1

2

2

1

16,000/N/A

0.8/N/A

ecs.e-c1m2.xlarge

4

8.0

0.4/Up to 3

1

2

6

1

16,000/N/A

0.8/N/A

ecs.e-c1m4.xlarge

4

16.0

0.8/Up to 4

1

2

6

1

16,000/N/A

0.8/N/A

ecs.e-c1m2.2xlarge

8

16.0

0.8/Up to 6

2

2

6

1

16,000/N/A

0.8/N/A

ecs.e-c1m4.2xlarge

8

32.0

1.2/Up to 6

2

2

6

1

16,000/N/A

0.8/N/A

Note
  • The ecs.e-c4m1.large, ecs.e-c2m1.large, ecs.e-c1m1.large, ecs.e-c1m2.large, and ecs.e-c1m4.large instance types have the following limits:

    • You cannot add secondary elastic network interfaces (ENIs) when you create an instance. You can add them after the instance is created.

    • To attach or detach a secondary ENI, the instance must be in the Stopped state.

  • The ecs.e-c4m1.large and ecs.e-c2m1.large instance types are available only in the following regions: China (Hong Kong), Singapore, Malaysia (Kuala Lumpur), Indonesia (Jakarta), Philippines (Manila), Thailand (Bangkok), Japan (Tokyo), South Korea (Seoul), UK (London), Germany (Frankfurt), US (Virginia), and US (Silicon Valley).

t6, burstable instance family

Features:

  • Baseline CPU performance with credit-based bursting

  • Better cost-performance ratio than t5

  • Processor: 2.5 GHz base / 3.2 GHz turbo, Intel Xeon Cascade Lake

  • Memory: DDR4

  • Storage: I/O optimized. Supports ESSDs, ESSD AutoPL disks, standard SSDs, and ultra disks.

  • Network: IPv4 and IPv6. VPC only.

Important

ESSDs at performance level (PL) 2 and PL3 cannot deliver peak performance on burstable instances. Use enterprise-grade instance types or ESSDs at lower PLs instead.

Use cases: Web application servers, lightweight applications and microservices, development and testing environments.

t6 instance types

Instance type

vCPUs

Memory (GiB)

Average baseline CPU performance

CPU credits/hour

Max CPU credit balance

Network baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

ecs.t6-c4m1.large

2

0.5

5%

6

144

0.08/up to 0.4

40,000

1

2

2

1

ecs.t6-c2m1.large

2

1.0

10%

12

288

0.08/up to 0.6

60,000

1

2

2

1

ecs.t6-c1m1.large

2

2.0

20%

24

576

0.08/up to 1

100,000

1

2

2

1

ecs.t6-c1m2.large

2

4.0

20%

24

576

0.08/up to 1

100,000

1

2

2

1

ecs.t6-c1m4.large

2

8.0

30%

36

864

0.08/up to 1

100,000

1

2

2

1

ecs.t6-c1m4.xlarge

4

16.0

40%

96

2,304

0.16/up to 2

200,000

1

2

6

1

ecs.t6-c1m4.2xlarge

8

32.0

40%

192

4,608

0.32/up to 4

400,000

2

2

6

1

Note
  • This family does not support adding secondary elastic network interfaces (ENIs) during instance creation. Attach one after the instance is created. The following types require the instance to be in the Stopped state to attach or detach a secondary ENI: ecs.t6-c1m1.large, ecs.t6-c1m2.large, ecs.t6-c1m4.large, ecs.t6-c2m1.large, ecs.t6-c4m1.large.

  • To check regional availability, see the Instance Types Available for Each Region page.

  • For metric definitions, see Instance type metrics.

t5, burstable instance family

Features:

  • Baseline CPU performance with credit-based bursting

  • Balanced compute, memory, and network resources

  • Processor: 2.5 GHz base, Intel Xeon

  • Memory: DDR4

  • Storage: Supports ultra disks and standard SSDs only

  • Network: IPv4 and IPv6. VPC only.

Use cases: Web application servers, lightweight applications and microservices, development and testing environments.

t5 instance types

Instance type

vCPUs

Memory (GiB)

Average baseline CPU performance

CPU credits/hour

Max CPU credit balance

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (pps)

NIC queues

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

ecs.t5-lc2m1.nano

1

0.5

20%

12

288

0.1

40,000

1

2

2

1

ecs.t5-lc1m1.small

1

1.0

20%

12

288

0.2

60,000

1

2

2

1

ecs.t5-lc1m2.small

1

2.0

20%

12

288

0.2

60,000

1

2

2

1

ecs.t5-lc1m2.large

2

4.0

20%

24

576

0.4

100,000

1

2

2

1

ecs.t5-lc1m4.large

2

8.0

20%

24

576

0.4

100,000

1

2

2

1

ecs.t5-c1m1.large

2

2.0

25%

30

720

0.5

100,000

1

2

2

1

ecs.t5-c1m2.large

2

4.0

25%

30

720

0.5

100,000

1

2

2

1

ecs.t5-c1m4.large

2

8.0

25%

30

720

0.5

100,000

1

2

2

1

ecs.t5-c1m1.xlarge

4

4.0

25%

60

1,440

0.8

200,000

1

2

6

1

ecs.t5-c1m2.xlarge

4

8.0

25%

60

1,440

0.8

200,000

1

2

6

1

ecs.t5-c1m4.xlarge

4

16.0

25%

60

1,440

0.8

200,000

1

2

6

1

ecs.t5-c1m1.2xlarge

8

8.0

25%

120

2,880

1.2

400,000

1

2

6

1

ecs.t5-c1m2.2xlarge

8

16.0

25%

120

2,880

1.2

400,000

1

2

6

1

ecs.t5-c1m4.2xlarge

8

32.0

25%

120

2,880

1.2

400,000

1

2

6

1

ecs.t5-c1m1.4xlarge

16

16.0

25%

240

5,760

1.2

600,000

2

2

6

1

ecs.t5-c1m2.4xlarge

16

32.0

25%

240

5,760

1.2

600,000

2

2

6

1

Note
  • This family does not support adding secondary ENIs during instance creation. Attach one after the instance is created. The following types require the instance to be in the Stopped state to attach or detach a secondary ENI: ecs.t5-lc2m1.nano, ecs.t5-c1m1.large, ecs.t5-c1m2.large, ecs.t5-c1m4.large, ecs.t5-lc1m1.small, ecs.t5-lc1m2.large, ecs.t5-lc1m2.small, ecs.t5-lc1m4.large.

  • To check regional availability, see the Instance Types Available for Each Region page.

  • For metric definitions, see Instance type metrics.

v5, CPU overprovisioned instance family

Features
  • You can create v5 instances only on dedicated hosts.
  • Compute:
    • Supports multiple CPU-to-memory ratios such as 1:1, 1:2, 1:4, and 1:8.
    • Uses 2.5 GHz Intel® Xeon® Platinum 8163 (Skylake) processors.
  • Storage:
    • Is an instance family in which all instances are I/O optimized.
    • Supports ESSDs, standard SSDs, and ultra disks.
  • Network:
    • Supports IPv6.
  • Suits the following scenarios:
    • Migration from offline virtualization environments to Alibaba Cloud
    • Services that generate low, medium, or burstable CPU loads
Instance types
Instance type vCPUs Memory (GiB) Bandwidth (Gbit/s) Packet forwarding rate (Kpps) NIC queues ENIs Private IP addresses per ENI
ecs.v5-c1m1.large 2 2.0 2.0 300 2 2 2
ecs.v5-c1m1.xlarge 4 4.0 2.0 300 2 2 6
ecs.v5-c1m1.2xlarge 8 8.0 3.0 400 2 3 6
ecs.v5-c1m1.3xlarge 12 12.0 3.0 400 4 3 6
ecs.v5-c1m1.4xlarge 16 16.0 4.0 500 4 4 6
ecs.v5-c1m1.8xlarge 32 32.0 4.0 500 8 4 6
ecs.v5-c1m2.large 2 4.0 2.0 300 2 2 2
ecs.v5-c1m2.xlarge 4 8.0 2.0 300 2 2 6
ecs.v5-c1m2.2xlarge 8 16.0 3.0 400 2 3 6
ecs.v5-c1m2.3xlarge 12 24.0 3.0 400 4 3 6
ecs.v5-c1m2.4xlarge 16 32.0 4.0 500 4 4 6
ecs.v5-c1m2.8xlarge 32 64.0 4.0 500 8 4 6
ecs.v5-c1m4.large 2 8.0 2.0 300 2 2 2
ecs.v5-c1m4.xlarge 4 16.0 2.0 300 2 2 6
ecs.v5-c1m4.2xlarge 8 32.0 3.0 400 2 3 6
ecs.v5-c1m4.3xlarge 12 48.0 3.0 400 4 3 6
ecs.v5-c1m4.4xlarge 16 64.0 4.0 500 4 4 6
ecs.v5-c1m4.8xlarge 32 128.0 4.0 500 8 4 6
ecs.v5-c1m8.large 2 16.0 2.0 300 2 2 2
ecs.v5-c1m8.xlarge 4 32.0 2.0 300 2 2 6
ecs.v5-c1m8.2xlarge 8 64.0 3.0 400 2 3 6
ecs.v5-c1m8.3xlarge 12 96.0 3.0 400 4 3 6
ecs.v5-c1m8.4xlarge 16 128.0 4.0 500 4 4 6
ecs.v5-c1m8.8xlarge 32 256.0 4.0 500 8 4 6
Note For more information about these specifications, see Instance family overview.

xn4, n4, mn4, and e4, previous-generation shared instance families

The xn4, n4, mn4, and e4 families provide the following features:

  • Multiple processor-to-memory ratios.

  • Processor: 2.5 GHz Intel® Xeon® processor.

  • Uses DDR4 memory.

  • I/O optimized instances.

  • Supports only IPv4.

Family

Features

vCPU-to-memory ratio

Scenarios

xn4

Shared basic instance

1:1

  • Frontend server for the web application

  • Light-load applications and microservices

  • Development, testing, and stress testing applications

n4

Shared compute-optimized instance

1:2

  • Websites and web applications

  • Development environments, build servers, code repositories, microservices, and test and staging environments

  • Lightweight enterprise applications

mn4

Shared general-purpose instance

1:4

  • Websites and web applications

  • Lightweight databases and caches

  • General applications and lightweight enterprise services

e4

Shared Memory Instances

1:8

  • Applications that require large amounts of memory

  • Lightweight databases and caches

Shared General Purpose xn4 includes the instance types and performance metrics listed in the following table:

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Packet forwarding rate (PPS) (10,000)

Multi-queue

ENIs

Private IPv4 addresses per ENI

ecs.xn4.small

1

1.0

0.5

5

1

2

2

Note
  • You cannot add secondary ENIs when you create an instance of this family. You can add them after the instance is created. To attach or detach a secondary ENI, an ecs.xn4.small instance must be in the Stopped state.

  • For more information about these specifications, see the "Instance type specifications" section in Overview of instance families. Packet forwarding rates vary significantly based on business scenarios. We recommend that you perform business stress tests on instances to choose appropriate instance types.

Shared compute-optimized n4 includes the instance types and performance metrics that are listed in the following table.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Packet forwarding rate (PPS) (10,000)

Multi-queue

ENIs

Private IPv4 addresses per ENI

ecs.n4.small

1

2.0

0.5

5

1

2

2

ecs.n4.large

2

4.0

0.5

10

1

2

2

ecs.n4.xlarge

4

8.0

0.8

15

1

2

6

ecs.n4.2xlarge

8

16.0

1.2

30

1

2

6

ecs.n4.4xlarge

16

32.0

2.5

40

1

2

6

ecs.n4.8xlarge

32

64.0

5.0

50

2

2

6

Note
  • You cannot add secondary ENIs when you create an instance of this family. You can add them after the instance is created. To attach or detach a secondary ENI, some instance types must be in the Stopped state. These instance types include ecs.n4.small and ecs.n4.large.

  • For more information about these specifications, see the "Instance type specifications" section in Overview of instance families. Packet forwarding rates vary significantly based on business scenarios. We recommend that you perform business stress tests on instances to choose appropriate instance types.

Shared general-purpose mn4 includes the instance types and performance metrics listed in the following table.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Packet forwarding rate (PPS) (10,000)

Multi-queue

ENIs

Private IPv4 addresses per ENI

ecs.mn4.small

1

4.0

0.5

5

1

2

2

ecs.mn4.large

2

8.0

0.5

10

1

2

2

ecs.mn4.xlarge

4

16.0

0.8

15

1

2

6

ecs.mn4.2xlarge

8

32.0

1.2

30

1

2

6

ecs.mn4.4xlarge

16

64.0

2.5

40

1

8

6

ecs.mn4.8xlarge

32

128.0

5

50

2

8

6

Note
  • You cannot add secondary ENIs when you create an instance of this family. You can add them after the instance is created. To attach or detach a secondary ENI, some instance types must be in the Stopped state. These instance types include ecs.mn4.small and ecs.mn4.large.

  • For more information about these specifications, see the "Instance type specifications" section in Overview of instance families. Packet forwarding rates vary significantly based on business scenarios. We recommend that you perform business stress tests on instances to choose appropriate instance types.

Shared memory-optimized e4 includes the instance types and performance metrics listed in the following table:

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Packet forwarding rate (PPS) (10,000)

Multi-Queue

ENIs

Private IPv4 addresses per ENI

ecs.e4.small

1

8.0

0.5

5

1

2

2

ecs.e4.large

2

16.0

0.5

10

1

2

2

ecs.e4.xlarge

4

32.0

0.8

15

1

2

6

ecs.e4.2xlarge

8

64.0

1.2

30

1

3

6

ecs.e4.4xlarge

16

128.0

2.5

40

1

8

6

Note
  • You cannot add secondary ENIs when you create an instance of this family. You can add them after the instance is created. To attach or detach a secondary ENI, some instance types must be in the Stopped state. These instance types include ecs.e4.small and ecs.e4.large.

  • For more information about these specifications, see the "Instance type specifications" section in Overview of instance families. Packet forwarding rates vary significantly based on business scenarios. We recommend that you perform business stress tests on instances to choose appropriate instance types.

Enterprise Arm-based instance family

g8y, general-purpose instance family

  • Introduction: g8y instances are equipped with Yitian 710, a proprietary ARM-based CPU developed by Alibaba Cloud. Built on the fourth-generation SHENLONG architecture, g8y instances provide stable and predictable ultra-high performance. They also use chip-level acceleration to provide significant improvements in storage and network performance, and computing stability.

  • Scenarios: Containers, microservices, web and application servers, video encoding and decoding, high-performance computing, and CPU-based machine learning.

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: 2.75 GHz Yitian 710 processor that provides stable computing performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • For small instance types, network bandwidth is burstable.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

The g8y instance family includes the instance types and metrics that are described in the following table.

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Maximum number of attached data disks

Baseline/Burst disk IOPS

Baseline/Burst disk bandwidth (Gbit/s)

ecs.g8y.small

1

4

1/10

0.5 million

Up to 0.25 million

1

2

3

3

5

10,000/Up to 110,000

1/Up to 10

ecs.g8y.large

2

8

2/10

0.9 million

Up to 0.25 million

2

3

6

6

8

20,000/Up to 110,000

1.5/Up to 10

ecs.g8y.xlarge

4

16

3/10

1 million

Up to 0.25 million

4

4

15

15

8

40,000/Up to 110,000

2/Up to 10

ecs.g8y.2xlarge

8

32

5/10

1.6 million

Up to 0.25 million

8

4

15

15

16

50,000/Up to 110,000

3/Up to 10

ecs.g8y.4xlarge

16

64

10/25

3 million

0.4 million

16

8

30

30

16

80,000/Up to 110,000

5/Up to 10

ecs.g8y.8xlarge

32

128

16/25

5 million

0.75 million

32

8

30

30

16

125,000

10

ecs.g8y.16xlarge

64

256

32 / None

10 million

1.5 million

64

8

30

30

32

250,000

16

ecs.g8y.32xlarge

128

512

64 / None

20 million

3 million

64

15

30

30

32

500,000

32

Note

To use ecs.g8y.32xlarge, you must submit a ticket to request access.

Compute-optimized instance family c8y

  • Introduction: These instances feature Alibaba Cloud's in-house Yitian 710 Arm-based processors and are built on the fourth-generation SHENLONG architecture. They deliver stable, predictable, and exceptionally high performance. They use chip-level fast-path acceleration to achieve order-of-magnitude improvements in storage and network performance and to enhance computing stability.

  • Use cases: Ideal for containers, microservices, websites, application servers, video encoding and decoding, high-performance computing (HPC), and CPU-based machine learning.

  • Compute:

    • The vCPU-to-memory ratio is 1:2.

    • Processor: 2.75 GHz Yitian 710 processors deliver stable computing performance.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For more information, see IPv6 communication.

    • Supports Elastic RDMA Interface (ERI). For instructions on how to use ERI, see Enable eRDMA on enterprise-level instances.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Achieves an exceptionally high packet forwarding rate (PPS).

    • Smaller instance types support burstable network bandwidth.

    • Network performance scales with the instance type.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

c8y includes the instance types and metrics that are described in the following table:

Instance type

vCPUs

Memory (GiB)

Network bandwidth (baseline/burst, Gbit/s)

Network PPS

Connections

NIC queues

ENIs

Private IPv4s per ENI

IPv6s per ENI

Max data disks

Disk IOPS (baseline/burst)

Disk bandwidth (baseline/burst, Gbit/s)

ecs.c8y.small

1

2

1/10

500,000

Up to 250,000

1

2

3

3

5

10,000/Up to 110,000

1/Up to 10

ecs.c8y.large

2

4

2/10

900,000

Up to 250,000

2

3

6

6

8

20,000/Up to 110,000

1.5/Up to 10

ecs.c8y.xlarge

4

8

3/10

1,000,000

Up to 250,000

4

4

15

15

8

40,000/Up to 110,000

2/Up to 10

ecs.c8y.2xlarge

8

16

5/10

1,600,000

Up to 250,000

8

4

15

15

16

50,000/Up to 110,000

3/Up to 10

ecs.c8y.4xlarge

16

32

10/25

3,000,000

400,000

16

8

30

30

16

80,000/Up to 110,000

5/Up to 10

ecs.c8y.8xlarge

32

64

16/25

5,000,000

750,000

32

8

30

30

16

125,000

10

ecs.c8y.16xlarge

64

128

32/None

10,000,000

1,500,000

64

8

30

30

32

250,000

16

ecs.c8y.32xlarge

128

256

64/None

20,000,000

3,000,000

64

15

30

30

32

500,000

32

Note

To request access to the ecs.c8y.32xlarge instance type, submit a ticket.

r8y memory-optimized instance family

  • Introduction: Built on the fourth-generation SHENLONG architecture and powered by Alibaba Cloud's in-house Yitian 710 Arm-based processors, this instance family delivers predictable and consistent ultra-high performance. It also uses chip-level fast-path acceleration, delivering an order-of-magnitude improvement in storage performance, network performance, and computing stability.

  • Scenarios: Containers, microservices, websites and application servers, video encoding and decoding, high-performance computing, and CPU-based machine learning.

  • Compute:

    • The vCPU-to-memory ratio is 1:8.

    • Processor: 2.75 GHz Yitian 710 processors, providing stable computing performance.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Supports Elastic RDMA Interface (ERI). To learn how to use ERI, see Enable eRDMA on enterprise-level instances.

    • Supports jumbo frames. For more information, see Jumbo Frames.

    • Provides an ultra-high packet forwarding rate (PPS).

    • Smaller instance types feature burstable network bandwidth.

    • Network performance scales with the instance type. Larger instance types provide higher network performance.

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

The following table lists the specifications for the r8y instance family.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

NIC queues

ENIs

Private IPv4s per ENI

IPv6s per ENI

Max data disks

Disk IOPS

Disk bandwidth (Gbit/s)

ecs.r8y.small

1

8

1/10

500,000

Up to 250,000

1

2

3

3

5

10,000/Up to 110,000

1/Up to 10

ecs.r8y.large

2

16

2/10

900,000

Up to 250,000

2

3

6

6

8

20,000/Up to 110,000

1.5/Up to 10

ecs.r8y.xlarge

4

32

3/10

1,000,000

Up to 250,000

4

4

15

15

8

40,000/Up to 110,000

2/Up to 10

ecs.r8y.2xlarge

8

64

5/10

1,600,000

Up to 250,000

8

4

15

15

16

50,000/Up to 110,000

3/Up to 10

ecs.r8y.4xlarge

16

128

10/25

3,000,000

400,000

16

8

30

30

16

80,000/Up to 110,000

5/Up to 10

ecs.r8y.8xlarge

32

256

16/25

5,000,000

750,000

32

8

30

30

16

125,000

10

ecs.r8y.16xlarge

64

512

32/N/A

10,000,000

1,500,000

64

8

30

30

32

250,000

16

ecs.r8y.32xlarge

128

1024

64/N/A

20,000,000

3,000,000

64

15

30

30

32

500,000

32

Note

To use the ecs.r8y.32xlarge instance type, submit a ticket.

g6r general-purpose instance family

  • Introduction: Built on the third-generation SHENLONG architecture, g6r instances provide stable and predictable ultra-high performance. They also use chip-level acceleration to provide significant improvements in storage and network performance, and computing stability.

  • Scenarios: Containers, microservices, test and development (such as DevOps), web and application servers, game servers, and CPU-based machine learning inference.

  • Compute:

    • Processor-to-memory ratio of 1:4.

    • Processor: 2.8 GHz Ampere® Altra® processor that provides stable computing performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances provide an ultra-high packet forwarding rate.

    • For small instance types, network bandwidth is burstable.

    • The network performance of an instance corresponds to its instance type. A larger instance type provides higher network performance.

The g6r instance family includes the instance types and metrics listed in the following table.

Instance type

vCPU

Memory (GiB)

Baseline/Burst network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.g6r.large

2

8

1/10

0.9 million

Up to 0.25 million

2

3

6

1

12,500

1

ecs.g6r.xlarge

4

16

1.5/10

1 million

Up to 0.25 million

4

4

15

1

20,000

1.5

ecs.g6r.2xlarge

8

32

2.5/10

1.6 million

Up to 0.25 million

8

4

15

1

30,000

2

ecs.g6r.4xlarge

16

64

5/10

2 million

0.3 million

8

8

30

1

60,000

3

ecs.g6r.8xlarge

32

128

8/10

3 million

0.6 million

16

7

30

1

75,000

4

ecs.g6r.16xlarge

64

256

16 / None

6 million

0.9 million

32

8

30

1

150,000

8

Compute-optimized instance family c6r

  • Introduction: Built on the third-generation SHENLONG architecture, c6r instances deliver stable and predictable ultra-high performance. They also use chip-level fast-path acceleration, achieving order-of-magnitude improvements in storage performance, network performance, and compute stability.

  • Use cases:

    • Containers and microservices

    • Development and testing (DevOps)

    • Websites and application servers

    • CPU-based machine learning inference

    • High-performance scientific and engineering applications

  • Compute:

    • The vCPU-to-memory ratio is 1:2.

    • Processor: 2.8 GHz Ampere® Altra® processors.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For more information about IPv6 communication, see IPv6 communication.

    • These instances deliver an ultra-high packet forwarding rate (PPS).

    • Smaller instance types support burstable network bandwidth.

    • Network performance scales with the instance type.

The following table lists the specifications of instance types in the c6r family.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (baseline/burst, Gbit/s)

Packet rate (PPS)

Connections (thousands)

NIC queues

ENIs

Private IPv4s per ENI

IPv6s per ENI

Disk baseline IOPS

Disk baseline bandwidth (Gbit/s)

ecs.c6r.large

2

4

1/10

900,000

Up to 250

2

3

6

1

12,500

1

ecs.c6r.xlarge

4

8

1/10

1,000,000

Up to 250

4

4

15

1

20,000

1.5

ecs.c6r.2xlarge

8

16

2/10

1,600,000

Up to 250

8

4

15

1

30,000

2

ecs.c6r.4xlarge

16

32

5/10

2,000,000

300

8

8

30

1

60,000

3

ecs.c6r.8xlarge

32

64

8/10

3,000,000

600

16

7

30

1

75,000

4

ecs.c6r.16xlarge

64

128

16/None

6,000,000

900

32

8

30

1

150,000

8

ECS bare metal instance families

GPU elastic bare metal server instance family ebmgn9g

Important

The ebmgn9g instance family is available by invitation only. To request access, submit a ticket.

  • Instance family introduction: The ebmgn9g instance family is Alibaba Cloud's ninth-generation line of cost-effective, full-featured GPU bare metal instances. Powered by the latest CIPU 2.0, these instances combine high-frequency CPUs, large memory capacity, and new Blackwell architecture professional GPUs. They deliver high-performance and cost-effective GPU cloud services for a wide range of accelerated workloads, such as training for autonomous driving and embodied intelligence, large model inference, film and animation rendering, and services for the metaverse and cloud gaming.

  • Use cases and features:

    • Autonomous driving/embodied intelligence:
      With 256 vCPUs running at an all-core frequency of up to 4.2 GHz and 2,304 GiB of memory, these instances handle the demanding data processing requirements for autonomous driving and embodied intelligence training.

    • Search and recommendation:
      The Blackwell GPU provides 126 TFLOPS of TF32 computing power. Each GPU is paired with 32 vCPUs and 153 GB/s of memory bandwidth, making it an optimal configuration for search and advertising services.

    • Large model inference:
      The new-generation GPU offers a significant performance leap over its predecessor, with video memory bandwidth increased to 1,344 GB/s. Support for FP4 computing power significantly improves inference performance and cost-effectiveness. The eight GPUs are interconnected via PCIe Gen5 with a bandwidth of 128 GB/s. This connection dramatically improves multi-card parallel inference efficiency.

    • Cloud gaming/rendering/metaverse:
      With a CPU frequency of up to 5 GHz, these instances are an excellent choice for 3D modeling. The GPU provides certified, workstation-grade graphics drivers with full OpenGL acceleration, making these instances the optimal choice for high-end film and animation development and CAD design.

  • Powered by the latest CIPU 2.0:

    The second-generation CIPU offers increased processing power, enhancing the performance of eRDMA, VPC, and cloud disk components. Bare metal instances are ideal for workloads that require direct access to physical resources or have requirements such as hardware-based licensing. They also support containers, including but not limited to Docker, Clear Containers, and PouchContainer.

  • Compute:

    • Powered by new Blackwell architecture professional GPUs:

      • Supports professional-grade OpenGL graphics processing.

      • Supports common acceleration features such as RTX and TensorRT, with newly added support for FP4 and PCIe Gen5 interconnect.

      • Utilizes a PCIe switch for interconnection, which improves NCCL performance by 36% compared to a direct-to-CPU connection. This can boost performance by up to 9% for large model inference with multi-card sharding.

    • Key GPU specifications:

      GPU architecture

      Video memory

      Compute performance

      Encode/decode engines

      Inter-GPU interconnect

      Acceleration APIs

      Blackwell

      • Capacity: 48 GB

      • Bandwidth: 1,344 GB/s

      • TF32: 126 TFLOPS

      • FP32: 52 TFLOPS

      • FP16/BF16: 266 TFLOPS

      • FP8/INT8: 533 TFLOPS

      • FP4: 970 TFLOPS

      • RT core: 196 TFLOPS

      • 3 × Video Encoder

      • 3 × Video Decoder

      • PCIe Gen5 x16: 128 GB/s

      • Supports P2P

      Supports DX12,

      OpenGL 4.6, Vulkan 1.3, CUDA 12.8, OpenCL 3.0, and DirectCompute

    • Processor: AMD Turin-C processors with a frequency range of 3.3 GHz to 5 GHz and an all-core turbo of up to 4.2 GHz.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Delivers network performance of up to 30 million Network PPS.

    • Supports eRDMA, which enables RDMA pass-through acceleration over VPC networks. eRDMA increases bandwidth to 360 Gbit/s and is ideal for training workloads in autonomous driving, embodied intelligence, computer vision (CV), and traditional models.

      Note

      For instructions on how to use eRDMA, see Enable eRDMA for an enterprise-level instance.

The following table lists the specifications for the ebmgn9g instance family.

Instance type

vCPU

Memory (GiB)

Video memory

Network bandwidth (Gbit/s)

Network PPS

Private IPv4 addresses

IPv6 addresses

Queues (primary/secondary)

ENIs

Max data disks

Cloud disk bandwidth (GB/s)

ecs.ebmgn9g.64xlarge

256

2304

48 GB × 8

360 (180 × 2)

30 million

30

30

64/16

38

33

8

Note

ebmgn9g instances require images that use the UEFI boot mode. If you use a custom image, ensure it supports the UEFI boot mode and that its boot mode property is set to UEFI. For more information, see Instance boot mode.

GPU elastic bare metal server instance family ebmgn9ge

Important

The ebmgn9ge instance family is available by invitation only. To request access, submit a ticket.

  • Introduction: The ebmgn9ge instance family is Alibaba Cloud's 9th generation of full-featured, cost-effective GPU bare metal instances. Powered by the latest CIPU 2.0, these instances combine high-frequency CPUs, large memory capacity, and new Blackwell architecture professional GPUs. They deliver high-performance and cost-effective GPU cloud services for a wide range of accelerated workloads, including autonomous driving and embodied intelligence training, large model inference, film and animation rendering, and metaverse and cloud gaming services.

  • Use cases and features:

    • Autonomous driving and embodied intelligence:
      With 256 vCPUs running at an all-core frequency of over 4.2 GHz and 2.3 TB of memory, these instances meet the demanding data processing needs of autonomous driving and embodied intelligence training.

    • Search and recommendation:
      The Blackwell GPU provides 126 TFLOPS of TF32 computing power. Each GPU is paired with 32 vCPUs and 153 GB/s of memory bandwidth, providing an optimal configuration for search and advertising services.

    • Large model inference:

      The ebmgn9ge instance family is specifically designed for large language models, offering 72 GB of GPU memory per GPU. The GPU memory bandwidth reaches 1,344 GB/s, delivering high-performance inference for LLM scenarios. Combined with the new FP4 compute architecture and 128 GB/s of PCIe Gen5 bandwidth, an instance can support parallel inference for models larger than 671 billion parameters across 8 GPUs.

    • Cloud gaming, rendering, and metaverse:
      With a CPU frequency of up to 5 GHz, these instances are well-suited for 3D modeling. The GPU natively supports graphics capabilities and provides certified workstation-grade graphics drivers with full OpenGL acceleration, making them ideal for high-end film and animation development and CAD design.

  • Powered by the latest CIPU 2.0:

    The 2nd generation CIPU delivers greater cloud processing power, enhancing performance for eRDMA, VPC, and EBS components. Bare metal instances allow workloads to directly access physical resources or meet requirements such as hardware-based licensing. They also support containers, such as Docker, Clear Containers, and Pouch.

  • Compute:

    • Powered by new Blackwell architecture professional GPUs:

      • Supports professional-grade OpenGL graphics processing.

      • Supports common acceleration features like RTX and TensorRT, and adds support for FP4 compute and the PCIe Gen5 interconnect.

      • Utilizes a PCIe switch for interconnection, which improves NCCL performance by 36% compared to a direct-to-CPU connection. This can boost performance by up to 9% for large model inference with multi-GPU sharding.

    • GPU key specifications:

      GPU architecture

      GPU memory

      Compute performance

      Video encode/decode engines

      Inter-GPU interconnect

      Accelerated APIs

      Blackwell

      • Capacity: 72 GB

      • Bandwidth: 1,344 GB/s

      • TF32: 126 TFLOPS

      • FP32: 52 TFLOPS

      • FP16/BF16: 266 TFLOPS

      • FP8/INT8: 533 TFLOPS

      • FP4: 970 TFLOPS

      • RT core: 196 TFLOPS

      • 3 × Video Encoder

      • 3 × Video Decoder

      • PCIe Gen5 x16: 128 GB/s

      • Supports P2P

      Supports DX12,

      OpenGL 4.6, Vulkan 1.3, CUDA 12.8, OpenCL 3.0, and DirectCompute

    • Processor: AMD Turin-C processors with a frequency range of 3.3 GHz to 5 GHz and an all-core turbo of up to 4.2 GHz.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 Communication.

    • Delivers network performance of up to 30 million network packets per second (PPS).

    • Supports ERI (Elastic RDMA Interface), which enables RDMA pass-through acceleration over VPC networks. ERI increases bandwidth to 360 Gbit/s and is ideal for training workloads in autonomous driving, embodied intelligence, computer vision (CV), and traditional models.

      Note

      For instructions on how to use ERI, see Enable eRDMA on enterprise-level instances.

The following table lists the instance types and specifications for the ebmgn9ge instance family.

Note

For a lower-cost version with less memory, you can select ebmgn9gc.

Instance type

vCPU

Memory (GiB)

GPU memory

Baseline network bandwidth (Gbit/s)

Network PPS

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Queue pairs (primary/secondary ENI)

ENIs

Max attachable data disks

Max cloud disk bandwidth (GB/s)

ecs.ebmgn9ge.64xlarge

256

2,304

72 GB × 8

360 (180 × 2)

30 million

30

30

64/16

38

33

8

Note

Instances of the ebmgn9ge instance family require an image that uses the UEFI boot mode. If you use a custom image, ensure that the image supports the UEFI boot mode and that its boot mode property is set to UEFI. For more information, see Instance Boot Mode.

GPU-accelerated elastic bare metal instance family ebmgn9gc

Important

The ebmgn9gc instance family is available by invitation only. To request access, submit a ticket.

  • Introduction: The ebmgn9gc instance family is Alibaba Cloud's 9th generation of full-featured, cost-effective GPU bare metal instances. Powered by the latest CIPU 2.0, these instances combine high-frequency CPUs, large-capacity memory, and new Blackwell architecture professional GPUs to deliver cost-effective GPU cloud services for a wide range of accelerated workloads, including autonomous driving and embodied intelligence training, large model inference, film and animation rendering, and metaverse and cloud gaming services.

  • Use cases and features:

    • Autonomous driving and embodied intelligence:
      Offers 256 vCPUs with a frequency range of 3.3 GHz to 5.0 GHz and an all-core turbo of up to 4.2 GHz. Paired with 1.5 TB of memory, these instances meet the data processing demands of autonomous driving and embodied intelligence training.

    • Search and recommendation:
      The equipped Blackwell GPUs provide 126 TFLOPS of high-performance TF32 compute. Each GPU is paired with 32 vCPUs and 153 GB/s of memory bandwidth, providing an optimal configuration for search and advertising services.

    • Large model inference:

      The ebmgn9gc instances are designed for large language models. With 72 GB of GPU memory and 1,344 GB/s of memory bandwidth per GPU, they deliver high-performance inference for LLM scenarios. Combined with the new FP4 compute architecture and 128 GB/s of PCIe Gen5 bandwidth, an 8-card instance can support parallel inference for models with over 671 billion parameters.

    • Cloud gaming, rendering, and metaverse:
      With a CPU frequency of up to 5.0 GHz, these instances are well-suited for 3D modeling. The GPUs natively support graphics capabilities and provide certified workstation-grade graphics drivers with full OpenGL acceleration, making them a good choice for high-end film and animation development and CAD design.

  • Powered by the latest CIPU 2.0:

    The 2nd generation CIPU offers higher cloud processing power and enhanced computing capabilities for eRDMA, VPC, and block storage. Bare metal instances allow workloads to directly access physical resources or meet requirements such as hardware-based licensing. They also support containers, such as Docker, Clear Container, and Pouch.

  • Compute:

    • Powered by new Blackwell architecture professional GPUs:

      • Supports professional-grade OpenGL graphics processing features.

      • Supports common acceleration features such as RTX and TensorRT, and adds support for FP4 and PCIe Gen5 interconnect.

      • Utilizes a PCIe switch for interconnection, which improves NCCL performance by 36% compared to a direct-to-CPU connection. This can boost performance by up to 9% for large model inference with multi-card sharding.

    • GPU key parameters:

      GPU architecture

      GPU memory

      Compute performance

      Video encoding/decoding

      GPU-to-GPU interconnect

      Acceleration APIs

      Blackwell

      • Capacity: 72 GB

      • Bandwidth: 1,344 GB/s

      • TF32: 126 TFLOPS

      • FP32: 52 TFLOPS

      • FP16/BF16: 266 TFLOPS

      • FP8/INT8: 533 TFLOPS

      • FP4: 970 TFLOPS

      • RT Core: 196 TFLOPS

      • 3 × Video Encoder

      • 3 × Video Decoder

      • PCIe Gen5 x16: 128 GB/s

      • Supports P2P

      Supports DX12,

      OpenGL 4.6, Vulkan 1.3, CUDA 12.8, OpenCL 3.0, and DirectCompute

    • Processor: AMD Turin-C processor with a frequency range of 3.3 GHz to 5.0 GHz and an all-core turbo of up to 4.2 GHz.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Delivers network performance of up to 30 million packets per second (PPS).

    • Supports elastic RDMA (eRDMA), which enables RDMA pass-through acceleration over VPC networks. eRDMA increases bandwidth to 360 Gbit/s and is ideal for training workloads in autonomous driving, embodied intelligence, computer vision (CV), and traditional models.

      Note

      For instructions on how to use eRDMA, see Enable eRDMA on enterprise-level instances.

The following table lists the instance types and specifications for the ebmgn9gc instance family.

Note

For a version with more memory, you can select ebmgn9ge.

Instance type

vCPU

Memory (GiB)

GPU memory

Network bandwidth (Gbit/s)

Network PPS

Private IPv4 addresses

IPv6 addresses

Multi-queue (primary/secondary)

Elastic network interfaces

Maximum data disks

Cloud disk bandwidth (GB/s)

ecs.ebmgn9gc.64xlarge

256

1536

72 GB × 8

360 (180 × 2)

30,000,000

30

30

64/16

38

33

8

Note

Instances in the ebmgn9gc instance family must be launched from an image that uses the UEFI boot mode. If you use a custom image, ensure that it supports the UEFI boot mode and its boot mode property is set to UEFI. For more information, see Instance boot mode.

GPU compute-optimized elastic bare metal server instance family ebmgn8v

This instance family is currently available only in select regions. To request access, contact your Alibaba Cloud sales representative.

  • Instance family overview: ebmgn8v is Alibaba Cloud's 8th-generation accelerated computing instance family (Elastic Bare Metal Server instance family), designed for AI model training and ultra-large parameter models. Each instance is a bare metal host equipped with eight GPU cards.

  • Use cases:

    • Cost-effective for multi-GPU parallel inference on large language models (LLMs) with more than 70 billion parameters.

    • Each GPU delivers 39.5 TFLOPS of FP32 compute, making it ideal for traditional AI model training and autonomous driving workloads.

    • NVLINK interconnection among the eight GPUs supports small- and medium-sized model training scenarios.

  • Key features and positioning:

    • High-speed, large-capacity GPU memory: Each GPU has 96 GB of HBM3 memory, delivering a memory bandwidth of 4 TB/s. This significantly accelerates model training and inference.

    • High inter-GPU bandwidth: The GPUs are interconnected via 900 GB/s NVLINK, providing significantly higher efficiency for multi-GPU training and inference compared to previous GPU generations.

    • Large model quantization technology: Supports FP8 compute, which optimizes performance for training and inference with large-scale parameters. This significantly increases computational speed and reduces GPU memory usage.

  • Compute:

    • Powered by the latest CIPU (Cloud Processing Unit) 1.0:

      • Decouples compute and storage, enabling flexible selection of storage resources. Compared to 7th-generation GPU instances, the inter-host bandwidth is increased to 160 Gbit/s for faster data transfer and processing.

      • Unlike traditional virtualized instances, the CIPU provides bare metal capabilities that support P2P communication between GPU instances.

    • Powered by 4th Generation Intel® Xeon® Scalable processors, providing 192 vCPUs with an all-core turbo frequency of up to 3.1 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Delivers ultra-high network performance of up to 30 million PPS.

    • Supports Elastic RDMA Interface (ERI), which enables RDMA pass-through acceleration over VPC networks. ERI increases bandwidth to 160 Gbit/s and is ideal for computer vision (CV) and traditional model training workloads.

      Note

      For instructions on how to use ERI, see Enable eRDMA on enterprise-level instances.

The following table describes the instance types and specifications of the ebmgn8v instance family.

Instance type

vCPU

Memory (GiB)

GPU memory

Baseline network bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Private IPv4s per ENI

IPv6s per ENI

Multi-queue

ENIs

Max data disks

Max cloud disk bandwidth (GB/s)

ecs.ebmgn8v.48xlarge

192

1,024

96 GB × 8

170 (85 × 2)

30 million

30

30

64

32

31

6

Note

The boot mode for images used by ebmgn8v instances must be UEFI. If you use a custom image, ensure that the image supports the UEFI boot mode and that its boot mode attribute is set to UEFI. For more information, see Instance boot modes.

GPU-accelerated ebmgn8ia instance family

This instance family is currently available only in select regions. To request access, contact your Alibaba Cloud sales representative.

  • Introduction: The ebmgn8ia, an elastic bare metal instance family, is Alibaba Cloud's eighth-generation accelerated computing instance family. It is designed for search and recommendation, simulation, and other GPU-intensive workloads that require a high vCPU-to-GPU ratio. Powered by the latest NVIDIA L20 GPUs, each instance is a bare metal server equipped with two high-frequency processors and four GPUs.

  • Features and use cases:

    • High frequency: These instances are equipped with two AMD EPYC™ Genoa 9T34 processors. Each processor has 64 physical cores, providing a total of 256 vCPUs per instance, with a frequency range of 3.4 GHz to 3.75 GHz. This significantly boosts single-core CPU performance, making it ideal for CAD modeling and accelerating preprocessing for CAE simulations.

    • Sparse resource ratio: Each GPU is paired with 64 vCPUs and 384 GiB of memory, delivering an average memory bandwidth of 230 GB/s per GPU. This configuration is suitable for GPU computing scenarios that require high I/O throughput, such as advertising, search, recommendation, traditional CAE simulations, and certain CPU-based rendering tasks in film production.

  • Powered by the latest CIPU 1.0:

    • Decouples compute and storage, allowing you to flexibly select storage resources. Compared to the previous generation, the inter-machine bandwidth of this instance type is increased to 160 Gbit/s for faster data transfer and processing.

    • The CIPU provides bare metal capabilities, enabling PCIe P2P communication between GPU instances, an improvement over traditional virtualized instances.

  • Compute:

    • Powered by new enterprise-class NVIDIA L20 GPUs:

      • Supports common acceleration features such as vGPU, RTX, and TensorRT.

      • Supports FP8 precision to improve computational efficiency.

    • NVIDIA L20 key specifications:

      GPU architecture

      GPU memory

      Compute performance

      Video encode/decode engines

      GPU interconnect

      NVIDIA Ada Lovelace

      • Capacity: 48 GB

      • Bandwidth: 864 GB/s

      • FP64: N/A

      • FP32: 59.3 TFLOPS

      • FP16/BF16: 119 TFLOPS

      • FP8/INT8: 237 TFLOPS

      • 3 × Video Encoder (+AV1)

      • 3 × Video Decoder

      • 4 × JPEG Decoder

      • PCIe interface: PCIe Gen4 x16

      • Bandwidth: 64 GB/s

    • Processor: AMD EPYC™ Genoa 9T34 processors with a frequency range of 3.4 GHz to 3.75 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Delivers ultra-high network performance of up to 30,000,000 PPS.

    • Supports Elastic RDMA Interface (ERI), which enables RDMA passthrough over VPC networks. ERI increases bandwidth to 160 Gbit/s and is ideal for training computer vision (CV) and traditional models.

      Note

      For instructions on using ERI, see Enable eRDMA on an enterprise-level instance.

The following table lists the instance types and specifications for the ebmgn8ia instance family.

Instance type

vCPU

Memory (GiB)

GPU

GPU memory

Baseline network bandwidth (Gbit/s)

Network PPS

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Queue number (Primary/Secondary ENI)

Elastic network interfaces

Max data disks

Max cloud disk bandwidth (GB/s)

ecs.ebmgn8ia.64xlarge

256

1536

L20 × 4

48 GB × 4

160 (80 × 2)

30,000,000

30

30

64/16

32

31

6

Note

Instances of the ebmgn8ia instance family require an image that uses the UEFI boot mode. If you use a custom image, it must support UEFI boot, and its boot mode property must be set to UEFI. For more information, see Instance boot mode.

GPU compute-optimized elastic bare metal family ebmgn8is

This instance family is currently available only in select regions outside the Chinese mainland. For inquiries, contact your Alibaba Cloud sales representative.

  • Instance family overview: The ebmgn8is is the 8th-generation accelerated computing instance family from Alibaba Cloud. As an elastic bare metal instance family, it is built for the growing demand in AI-generated workloads. Powered by the latest NVIDIA L20 GPUs, each instance is a bare metal host with eight GPU cards.

  • Features and positioning:

    • Graphics processing: These instances are powered by high-frequency 4th Generation Intel® Xeon® Scalable processors. They provide ample CPU computing power for 3D modeling, ensuring smoother graphics rendering and design.

    • Inference tasks: Powered by the new NVIDIA L20 GPUs, where each card provides 48 GB of memory to accelerate inference tasks. These instances support the FP8 floating-point format. Combined with Alibaba Cloud Container Service for Kubernetes (ACK), they flexibly support inference on various AIGC models and are especially suitable for inference tasks on large language models (LLMs) with up to 70 billion parameters.

    • Training tasks: These instances offer cost-effective computing power, delivering double the FP32 compute performance compared to 7th-generation inference instances. They are particularly suitable for training computer vision (CV) models that use FP32, and for training other small to medium-sized models.

  • Use cases:

    • Use GRID images with GRID graphics drivers from the Cloud Marketplace to enable OpenGL and Direct3D graphics capabilities. This provides workstation-grade graphics processing for workloads such as animation, film and television special effects, and rendering.

    • Combined with the containerized management capabilities of ACK, you can more efficiently and cost-effectively support AIGC graphics generation and large language model (LLM) inference (up to 130 billion parameters).

    • Other general AI scenarios, such as image recognition and speech recognition.

  • Powered by the latest CIPU 1.0:

    • Features decoupled compute and storage, which allows you to flexibly select the required storage resources. Compared with the previous generation, the inter-instance bandwidth of this instance family is increased to 160 Gbit/s to enable faster data transfer and processing.

    • CIPU provides bare metal capabilities that enable PCIe P2P communication between GPUs, which is not supported by traditional virtualized instances.

  • Compute:

    • Powered by new enterprise-class NVIDIA L20 GPUs:

      • Supports common acceleration features such as vGPU, RTX, and TensorRT.

      • Uses a PCIe switch for interconnection. Compared with a direct-to-CPU connection, this design improves NCCL performance by 36% and can increase inference performance by up to 9% when you run sharded inference for large models across multiple cards.

    • Key NVIDIA L20 parameters:

      GPU architecture

      GPU memory

      Compute performance

      Video encoding/decoding

      Inter-GPU interconnect

      NVIDIA Ada Lovelace

      • Capacity: 48 GB

      • Bandwidth: 864 GB/s

      • FP64: N/A

      • FP32: 59.3 TFLOPS

      • FP16/BF16: 119 TFLOPS

      • FP8/INT8: 237 TFLOPS

      • 3 × Video Encoder (+AV1)

      • 3 × Video Decoder

      • 4 × JPEG Decoder

      • PCIe interface: PCIe Gen4 x16

      • Bandwidth: 64 GB/s

    • Processor: Intel® Xeon® Scalable processors (SPR) with a 3.4 GHz base frequency and an all-core turbo frequency of up to 3.9 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Delivers ultra-high network performance with a packet forwarding rate of 30 million PPS.

    • Supports ERI (Elastic RDMA Interface), which enables RDMA passthrough acceleration in a VPC network. ERI increases bandwidth to 160 Gbit/s and is ideal for CV and traditional model training workloads.

The following table describes the instance types and specifications of the ebmgn8is instance family.

Instance type

vCPUs

Memory (GiB)

GPU

GPU memory

Base bandwidth (Gbit/s)

Packet forwarding rate (PPS)

Private IPv4 addresses

IPv6 addresses

Queue pairs (primary/secondary ENI)

ENIs

Attachable data disks

Cloud disk bandwidth (GB/s)

ecs.ebmgn8is.32xlarge

128

1024

L20 × 8

48 GB × 8

160 (80 × 2)

30,000,000

30

30

64/16

32

31

6

Note

The image boot mode for ebmgn8is instance types must be UEFI. If you use a custom image, you must ensure that it supports UEFI and that its boot mode attribute is set to UEFI. For more information, see Instance boot modes.

Ebmgn7e instance family

  • Instance family description: The ebmgn7e is an instance family built on the X-Dragon architecture. It combines powerful hardware performance with software-defined flexibility and elasticity.

  • Use cases:

    • Deep learning training and development.

    • HPC-accelerated computing and simulations.

    Important

    When you run AI training workloads with high communication loads, such as those involving Transformer models, you must enable NVLink for GPU-to-GPU communication. Otherwise, large-scale data transfers over the PCIe link may cause unexpected failures and data corruption. If you are unsure about the communication link topology for your training workload, submit a ticket for support from Alibaba Cloud technical experts.

  • Compute:

    • Processor: Features an Intel® Xeon® Scalable processor with a base frequency of 2.9 GHz, an all-core turbo frequency of 3.5 GHz, and support for PCIe 4.0.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Achieves a packet forwarding rate of up to 24 million PPS.

The following table lists the instance types and specifications for the ebmgn7e instance family.

Instance type

vCPU

Memory (GiB)

GPU memory

Network bandwidth (Gbit/s)

Network PPS

Multi-queue (primary/secondary)

ENIs

Private IPv4 addresses

IPv6 addresses

ecs.ebmgn7e.32xlarge

128

1024

80 GB × 8

64

24 million

32/12

32

10

1

After starting an ebmgn7e instance, you must check if the Multi-Instance GPU (MIG) feature is enabled or disabled because its default state is not guaranteed. For more information about MIG, see NVIDIA Multi-Instance GPU User Guide.

The following table describes whether ebmgn7e instances support the MIG feature.

Instance type

MIG feature support

Description

ecs.ebmgn7e.32xlarge

Yes

You can enable the MIG feature on ebmgn7e bare metal instances.

GPU-accelerated ebmgn7i instance family

  • Introduction: The ebmgn7i is an instance family built on the X-Dragon architecture. These instances provide software-defined hardware computing, delivering both elasticity and powerful performance.

  • Use cases:

    • Equipped with high-performance CPUs, memory, and GPUs, these instances can handle high-volume concurrent AI inference workloads, making them ideal for image, speech, and behavior recognition.

    • RTX support, combined with high-frequency CPUs, delivers high-performance 3D graphics virtualization ideal for intensive graphics processing workloads like remote graphic design and cloud gaming.

    • RTX support, combined with high network and cloud disk bandwidth, makes these instances ideal for building high-performance rendering farms.

    • With multiple GPUs and high network bandwidth, these instances support small-scale deep learning training workloads.

  • Compute:

    • Powered by NVIDIA A10 GPUs:

      • Innovative Ampere architecture.

      • Supports common acceleration features such as vGPU, RTX, and TensorRT.

    • Processor: Intel® Xeon® Scalable processors (Ice Lake) with a 2.9 GHz base frequency and a 3.5 GHz all-core turbo frequency.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Delivers ultra-high network performance with up to 24 million network PPS.

The following table lists the instance types and specifications for the ebmgn7i instance family.

Instance type

vCPU

Memory (GiB)

GPU

GPU memory

Network bandwidth (Gbit/s)

Network PPS

Multi-queue

Elastic network interfaces

Private IPv4 addresses

IPv6 addresses

ecs.ebmgn7i.32xlarge

128

768

NVIDIA A10 × 4

24 GB × 4

64

24 million

32

32

10

1

Gpu-accelerated Elastic Bare Metal Server instance family ebmgn7

  • Introduction: Based on the X-Dragon Architecture, the ebmgn7 instance type family features software-defined hardware, offering both flexibility and powerful computing performance.

  • Use cases:

    • Deep learning training workloads, such as image classification, autonomous driving, and speech recognition.

    • GPU-intensive scientific computing, such as computational fluid dynamics (CFD), computational finance, molecular dynamics, and environmental analysis.

  • Compute:

    • Processor: 2.5 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processors.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Network performance scales with the instance type.

The following table lists the instance types and specifications for the ebmgn7 instance type family.

Instance type

vCPU

Memory (GiB)

GPU memory

Baseline bandwidth (Gbit/s)

PPS

Multi-queue

ENIs

IPv4s per ENI

IPv6s per ENI

ecs.ebmgn7.26xlarge

104

768

40 GB × 8

30

18,000,000

16

15

10

1

Ebmgn6e GPU-accelerated instance family

  • Introduction:

    • The ebmgn6e instance family, built on the X-Dragon architecture, delivers software-defined hardware computing that combines flexibility, elasticity, and powerful performance.

    • These instances are powered by NVIDIA V100 (32 GB NVLink) GPU compute cards.

    • The GPU accelerator is the V100 (SXM2 package), which has the following features:

      • Innovative Volta architecture.

      • 32 GB of HBM2 gpu memory per GPU (with a gpu memory bandwidth of 900 GB/s).

      • 5,120 CUDA Cores per GPU.

      • 640 Tensor Cores per GPU.

      • Each GPU supports six bidirectional NVLink links. Each unidirectional link provides 25 GB/s of bandwidth, totaling 300 GB/s (6 × 25 × 2).

  • Use cases:

    • Deep learning, such as training and inference for AI algorithms like image classification, autonomous driving, and speech recognition.

    • Scientific computing, such as computational fluid dynamics, computational finance, molecular dynamics, and environmental analysis.

  • Compute:

    • vCPU-to-memory ratio of 1:8.

    • Processors: Intel® Xeon® Platinum 8163 (Skylake) with a base frequency of 2.5 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Network performance scales with the instance type.

The following table lists the instance types and specifications for the ebmgn6e instance family.

Instance type

vCPU

Memory (GiB)

GPU

GPU memory

Baseline bandwidth (Gbit/s)

PPS

Multi-queue

ENIs

Private IPv4 addresses

IPv6 addresses

ecs.ebmgn6e.24xlarge

96

768

NVIDIA V100 × 8

32 GB × 8

32

4,800,000

16

15

10

1

Ebmgn6v instance family

  • Instance family description:

    • The ebmgn6v instance family is built on the X-Dragon architecture. These instances offer high-performance, software-defined hardware computing with cloud elasticity.

    • Powered by NVIDIA V100 GPUs.

    • The V100 (SXM2 package) GPU accelerator features the following:

      • Innovative Volta architecture

      • 16 GB of HBM2 GPU memory per GPU, with a memory bandwidth of 900 GB/s

      • 5,120 CUDA Cores per GPU

      • 640 Tensor Cores per GPU

      • Each GPU supports six NVLink bidirectional links, with a unidirectional bandwidth of 25 GB/s per link, for a total bandwidth of 300 GB/s (6 × 25 × 2 = 300 GB/s)

  • Use cases:

    • Deep learning training and inference workloads, such as image classification, autonomous driving, and speech recognition

    • Scientific computing workloads, such as computational fluid dynamics, computational finance, molecular dynamics, and environmental analysis

  • Compute:

    • vCPU-to-memory ratio of 1:4.

    • Processor: Intel ® Xeon ® Platinum 8163 (Skylake) with a 2.5 GHz base frequency.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Network performance scales with the instance type.

The following table lists the instance types and specifications for the ebmgn6v instance family.

Instance type

vCPU

Memory (GiB)

GPU

GPU memory

Network bandwidth (Gbit/s)

Network PPS

Multi-queue

ENIs

IPv4s per ENI

IPv6s per ENI

ecs.ebmgn6v.24xlarge

96

384

NVIDIA V100 × 8

16 GB × 8

30

4,500,000

8

32

10

1

GPU compute instance family ebmgn6i

  • Introduction

    • ebmgn6i is an instance type family that is built on the X-Dragon Architecture, delivers software-defined hardware computing, and combines flexible elasticity with powerful performance.

    • These instances are powered by NVIDIA T4 GPU accelerators with the following features:

      • Innovative Turing architecture

      • 16 GB of GPU memory per GPU with a memory bandwidth of 320 GB/s

      • 2,560 CUDA Cores per GPU

      • Up to 320 Turing Tensor Cores per GPU

      • Variable-precision Tensor Cores deliver 65 TFLOPS of FP16 performance, 130 INT8 TOPS, and 260 INT4 TOPS.

  • Use cases

    • AI (DL/ML) inference, including applications such as computer vision, speech recognition, speech synthesis, NLP, machine translation, and recommendation systems.

    • Real-time cloud rendering for cloud gaming.

    • Real-time cloud rendering for AR/VR.

    • Compute-intensive graphics workloads or graphics workstations.

    • GPU-accelerated databases.

    • High-performance computing.

  • Compute

    • vCPU-to-memory ratio of 1:4.

    • Processor: Intel® Xeon® Platinum 8163 (Skylake) with a base frequency of 2.5 GHz.

  • Storage

  • Network

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Network performance scales with the instance type.

The following table lists the instance types and specifications for the ebmgn6i instance family.

Instance type

vCPU

Memory (GiB)

GPU

GPU memory

Network bandwidth (Gbit/s)

Forwarding rate (PPS)

Multiqueue

ENIs

Private IPv4 addresses

IPv6 addresses

ecs.ebmgn6i.24xlarge

96

384

NVIDIA T4 × 4

16 GB × 4

30

4.5 million

8

32

10

1

Compute-optimized elastic bare metal instance family ebmc9ae

  • Instance family overview: Built on Alibaba Cloud's new CIPU architecture and powered by the latest AMD EPYC Turin processors, this instance family provides stable compute performance, a powerful I/O engine, and chip-level security hardening.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High packet-rate networking scenarios, such as live comment streaming and telecom traffic forwarding.

    • Web front-end servers.

    • Massively multiplayer online (MMO) game front ends.

    • Data analytics, batch computing, and video encoding.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of 1:2.

    • Processor: AMD EPYC Turin processors with a maximum turbo frequency of 3.7 GHz. These instances use physical cores to deliver stable compute performance.

  • Storage:

  • Network:

The following table lists the instance types and specifications of the ebmc9ae instance family:

Instance type

vCPU

Memory (GiB)

Baseline/burst network bandwidth (Gbit/s)

Network PPS

Connections

Multi-queue

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline cloud disk IOPS

Baseline/burst cloud disk bandwidth (Gbit/s)

ecs.ebmc9ae.48xlarge

192

384

100/None

30,000,000

6,000,000

64

38

50

50

600,000

50/None

Compute-optimized ebmc9i instance family

  • Introduction: Built on Alibaba Cloud's new CIPU architecture and powered by Intel® Xeon® 6 processors with P-cores, these instances deliver stable compute performance, a powerful I/O engine, and chip-level security hardening.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Third-party hypervisor compatibility for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High packet-rate networking scenarios, such as live comment streaming and telecom traffic forwarding.

    • Web front-end servers.

    • Massively multiplayer online (MMO) game front ends.

    • Data analytics, batch computing, and video encoding.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of 1:2.

    • Processor: Powered by Intel® Xeon® Granite Rapids processors with a base frequency of 3.2 GHz and an all-core turbo frequency of 3.6 GHz.

  • Storage:

  • Network:

The following table lists the instance types and specifications for the ebmc9i instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

Multi-queue

ENIs

Private IPv4s/ENI

IPv6s/ENI

Disk IOPS

Disk bandwidth (Gbit/s)

ecs.ebmc9i.48xlarge

192

384

64/None

20,000,000

6,000,000

64 (primary network interface)/16 (secondary network interface)

32

50

50

500,000/800,000

32/None

General-purpose ebmc8a instance family

  • Instance family description: Built on the new Alibaba Cloud CIPU architecture and powered by AMD EPYC Genoa processors, this instance family delivers stable compute performance, a powerful I/O engine, and chip-level security hardening.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High packet-rate networking scenarios, such as live comment streaming and telecom traffic forwarding.

    • Web front-end servers.

    • Massively multiplayer online (MMO) game front ends.

    • Data analytics, batch processing, and video encoding.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of 1:2.

    • Processor: AMD EPYC Genoa processors with a max boost frequency of up to 3.7 GHz, delivering stable compute performance.

  • Storage:

  • Network:

The ebmc8a instance family includes the following instance types:

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

Multi-queue

ENIs

Private IPv4s per ENI

IPv6s per ENI

Cloud disk IOPS

Cloud disk bandwidth (Gbit/s)

ecs.ebmc8a.48xlarge

192

384

64/None

18,000,000

3,000,000

64

38

30

30

500,000/None

32/None

Compute-optimized elastic bare metal instance family ebmc8y

  • Instance family introduction: Built on the new Alibaba Cloud CIPU architecture, these instances deliver stable computing power, a robust I/O engine, and dedicated hardware with physical isolation.

  • Scenarios:

    • Workloads that require direct access to physical resources or have hardware-bound licensing requirements.

    • Compatibility with third-party hypervisors for hybrid and multi-cloud deployments.

    • Containers, such as Docker, Clear Containers, and Pouch.

    • Video on demand and live streaming.

    • Enterprise applications of all types and sizes.

    • Web and application servers.

    • Data analytics and computing.

    • High-performance scientific and engineering applications.

  • Compute:

    • These instances use Alibaba Cloud's proprietary Yitian 710 Arm architecture processors, which have a base frequency of at least 2.75 GHz. They do not use hyper-threading, ensuring stable computing performance.

  • Storage:

  • Network:

The following table lists the instance types and specifications for the ebmc8y instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

Queues

ENIs

Private IPv4 addresses

IPv6 addresses

Cloud disk IOPS

Cloud disk bandwidth (Gbit/s)

ecs.ebmc8y.32xlarge

128

256

64/None

20,000,000

3,000,000

64 (primary)/32 (secondary)

38

30

30

500,000/None

32/None

Compute-optimized elastic bare metal server instance family ebmc8i

  • Instance family overview: These instances are built on the new Alibaba Cloud CIPU architecture. They provide stable computing power, a powerful I/O engine, dedicated hardware, and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or have hardware-bound licensing requirements.

    • Workloads that are compatible with third-party hypervisors and require hybrid cloud and multi-cloud deployment.

    • Containers, such as Docker, Clear Container, and Pouch.

    • Workloads with high packet rates, such as live comment streaming and telecom service forwarding.

    • Web front-end servers.

    • Massively multiplayer online (MMO) game front ends.

    • Data analytics, batch computing, and video encoding.

    • High-performance scientific and engineering applications.

  • Compute:

    • Processor: Powered by Intel® Xeon® Emerald Rapids or Intel® Xeon® Sapphire Rapids processors with a base frequency of at least 2.7 GHz and an all-core turbo frequency of 3.2 GHz.

      Note

      When you purchase an instance of this family, the system randomly assigns one of the preceding processors. You cannot manually select a processor.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

    • For information about operating system compatibility, see Intel instance types and operating system compatibility.

  • Storage:

  • Network:

The following table lists the instance types and specifications for the ebmc8i instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Packet rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4 addresses

IPv6 addresses

Cloud disk IOPS

Cloud disk bandwidth (Gbit/s)

ecs.ebmc8i.48xlarge

192

512

100/None

30,000,000

4,000,000

64 (primary) / 16 (secondary)

72

30

30

1,000,000/None

48/None

Ebmc7 compute-optimized instance family

  • Introduction:

    • Built on the third-generation X-Dragon Architecture, these instances use chip-level acceleration and fast-path technology to provide stable, predictable, and ultra-high compute, storage, and network performance.

    • Dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or have hardware-based licensing requirements.

    • Hybrid cloud and multi-cloud deployments that use third-party hypervisors.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High packet-rate networking scenarios, such as live comment streaming and telecom traffic forwarding.

    • Web front-end servers.

    • Massively multiplayer online (MMO) game front ends.

    • Data analytics, batch computing, and video encoding.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of 1:2.

    • Processor: Intel® Xeon® Platinum 8369B (Ice Lake) with a 2.9 GHz base frequency and 3.5 GHz all-core turbo frequency.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Ultra-high network performance of up to 24 million network PPS.

The following table lists the specifications of the instance types in the ebmc7 instance family.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Network PPS

Connections

Elastic network interfaces

Private IPv4 per ENI

IPv6 per ENI

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.ebmc7.32xlarge

128

256

64

24,000,000

2,400,000

32

20

20

600,000

32

Compute-optimized EBMS instance family ebmc7a

  • Introduction:

    • Built on the third-generation X-Dragon Architecture, these instances use chip-level acceleration and fast-path technology to deliver stable, predictable, and ultra-high compute, storage, and network performance.

    • They provide dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • Video encoding/decoding and rendering.

    • Data analytics and computing.

  • Compute:

    • The vCPU-to-memory ratio is 1:2.

    • Processor: AMD EPYCTM MILAN processors with a base frequency of 2.55 GHz and a single-core turbo frequency of up to 3.5 GHz, delivering consistent compute performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Achieves network performance of up to 24 million PPS.

The following table lists the instance types and specifications for the ebmc7a instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

Multi-queue

ENIs

Private IPv4 addresses/ENI

IPv6 addresses/ENI

Baseline IOPS

Disk bandwidth (Gbit/s)

ecs.ebmc7a.64xlarge

256

512

64

24,000,000

4,000,000

32

31

15

1

600,000

32

Note
  • Instances of this type must use an image with the UEFI boot mode. If you use a custom image, ensure its boot mode is set to UEFI. For instructions on how to set the instance boot mode, see Instance Boot Mode.

  • Ubuntu 18 and Debian 9 kernels do not support AMD EPYC MILAN processors. Therefore, do not use Ubuntu 18 or Debian 9 images to create instances of this instance type. Otherwise, the instances will fail to start.

Compute-optimized ECS bare metal instance family ebmc6me

  • Introduction: Offers dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or have hardware-bound licensing requirements.

    • Third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • Video encoding/decoding and rendering.

    • Massively multiplayer online (MMO) game front ends.

    • High-performance scientific and engineering applications.

  • Compute:

    • Processor-to-memory ratio of approximately 1:3.

    • Processor: Intel® Xeon® Gold 5218 (Cascade Lake) with a base frequency of 2.3 GHz and a turbo frequency of 3.9 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • High network performance with a packet forwarding rate of up to 6 million PPS.

This table lists the instance types and specifications for the ebmc6me instance family.

Instance type

vCPUs

Memory (GiB)

Network bandwidth (Gbit/s)

Packet rate (PPS)

Connections

ENIs

Private IPv4 addresses / ENI

IPv6 addresses / ENI

Disk IOPS

Disk bandwidth (Gbit/s)

ecs.ebmc6me.16xlarge

64

192

32

6,000,000

1,800,000

32

10

1

200,000

16

Compute-optimized ebmc6a instance family

This instance family is available by invitation only. To request access, submit a ticket.

  • Overview:

    • Built on the third-generation X-Dragon architecture, these instances use chip-level fast-path acceleration to provide stable, predictable, and high-performance compute, storage, and networking.

    • They offer dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or have hardware-bound license requirements.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployment.

    • Containers, such as Docker, Clear Container, and Pouch.

    • Video encoding, decoding, and rendering.

    • Data analytics and compute.

  • Compute:

    • vCPU-to-memory ratio of 1:2.

    • Processor: 2.6 GHz AMD EPYCTM ROME processors with a 3.3 GHz turbo frequency, ensuring stable computing performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Ultra-high network performance of up to 24 million PPS.

The following table lists the instance types and specifications of the ebmc6a instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Packet rate (PPS)

Multi-queue

ENIs

Private IPv4 addresses

IPv6 addresses

Cloud disk IOPS

Cloud disk bandwidth (Gbit/s)

ecs.ebmc6a.64xlarge

256

512

64

24,000,000

32

31

10

1

600,000

32

Note

ECS instances of this instance family require images that use UEFI. If you use a custom image, you must set its boot mode to UEFI. For details on setting the image boot mode, see Instance boot mode.

Compute-optimized (balanced-enhanced) ebmc6e instance family

  • Instance family description:

    • Built on the third-generation X-Dragon Architecture, these instances use chip-level acceleration and fast-path technology to provide stable, predictable, and exceptionally high compute, storage, and network performance.

    • They provide dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Running third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High packet-rate scenarios, such as live comment streaming and telecom service forwarding.

    • Web front-end servers.

    • Massively multiplayer online (MMO) game front ends.

    • Data analytics, batch computing, and video encoding.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of approximately 1:2.

    • Processor: Intel ® Xeon ® Platinum 8269CY (Cascade Lake) with a base frequency of 2.5 GHz and an all-core turbo frequency of 3.2 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Provides network performance of up to 24 million network PPS.

The following table lists the instance types and specifications for the ebmc6e instance family.

Instance type

vCPUs

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Network PPS

Connections

Elastic network interfaces

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline cloud disk IOPS

Baseline cloud disk bandwidth (Gbit/s)

ecs.ebmc6e.26xlarge

104

192

32

24,000,000

1,800,000

32

10

1

480,000

16

Compute-optimized elastic bare metal instance family ebmc6

  • Instance family description: Offers dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • Video encoding/decoding and rendering.

    • Massively multiplayer online (MMO) game front ends.

    • High-performance scientific and engineering applications.

  • Compute:

    • Processor-to-memory ratio of approximately 1:2.

    • Processor: Intel® Xeon® Platinum 8269CY (Cascade Lake) with a base frequency of 2.5 GHz and an all-core turbo frequency of 3.2 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Offers a high network packet forwarding rate of up to 6 million PPS.

The following table lists the instance types and specifications for the ebmc6 instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

Elastic network interfaces

Private IPv4 addresses

IPv6 addresses

Baseline IOPS

Cloud disk bandwidth (Gbit/s)

ecs.ebmc6.26xlarge

104

192

32

6,000,000

1,800,000

32

20

1

200,000

16

General-purpose Elastic Bare Metal Server instance family ebmg9ae

  • Instance family overview: Built on the new Alibaba Cloud CIPU architecture and powered by the latest AMD EPYC Turin processors, this instance family delivers stable compute performance, a powerful I/O engine, and chip-level security hardening.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licensing.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Containers, and Pouch.

    • High packet-rate networking scenarios, such as live comment streaming and telecom traffic forwarding.

    • Enterprise applications of all types and sizes.

    • Web and application servers.

    • Game servers.

    • Small to medium-sized database systems, caches, and search clusters.

    • Data analytics and computing.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of 1:4.

    • Processor: AMD EPYC Turin processor with a max turbo frequency of 3.7 GHz. Its physical core design ensures stable compute performance.

  • Storage:

  • Network:

This table lists the specifications for the ebmg9ae instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

Multi-queue

ENIs

Private IPv4 addresses

IPv6 addresses

Baseline IOPS

Disk bandwidth (Gbit/s)

ecs.ebmg9ae.48xlarge

192

768

100/None

30,000,000

6,000,000

64

38

50

50

600,000

50/None

General-purpose ECS Bare Metal Instance family ebmg9a

  • Instance family overview: These instances are built on Alibaba Cloud's new CIPU architecture and powered by the latest AMD EPYC Turin processors. They deliver consistent compute performance, a powerful I/O engine, and chip-level security enhancements.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High packet-rate networking scenarios, such as live comment streaming and telecom traffic forwarding.

    • Enterprise applications of all types and sizes.

    • Website and application servers.

    • Game servers.

    • Small to medium-sized database systems, caches, and search clusters.

    • Data analytics and computing.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of 1:4.

    • Processor: AMD EPYC Turin with a max turbo frequency of up to 4.1 GHz, delivering stable compute performance.

  • Storage:

  • Network:

The following table lists the instance types and specifications for the ebmg9a instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

Multi-queue

ENIs

Private IPv4 addresses

IPv6 addresses

Cloud disk IOPS

Cloud disk bandwidth (Gbit/s)

ecs.ebmg9a.64xlarge

256

1152

100/None

30,000,000

6,000,000

64

38

50

50

500,000

64/None

General-purpose elastic bare metal server family ebmg9i

  • Instance family introduction: Built on Alibaba Cloud's new CIPU architecture and powered by Intel® Xeon® 6 processors with P-cores, these instances provide stable computing power, a more powerful I/O engine, and chip-level security hardening.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Compatible with third-party hypervisors for hybrid cloud and multi-cloud deployment.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High packet-rate networking scenarios, such as live comment streaming and telecom traffic forwarding.

    • Enterprise applications of all types and sizes.

    • Web and application servers.

    • Game servers.

    • Small to medium-sized database systems, caches, and search clusters.

    • Data analytics and computation.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of 1:4.

    • Processor: Intel® Xeon® Granite Rapids with a base frequency of 3.2 GHz and an all-core turbo frequency of 3.6 GHz.

  • Storage:

    • These are I/O optimized instances.

    • These instances support the NVMe protocol. For more information, see Overview of the NVMe protocol.

    • Supported cloud disk types: ESSD cloud disks, ESSD AutoPL cloud disks, and ESSD Zone-Redundant cloud disks. For details, see Block storage overview.

  • Network:

The following table lists the instance types and specifications for the ebmg9i instance type family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Packet rate (PPS)

Connections

Multi-queue

ENIs

Private IPv4s (per ENI)

IPv6s (per ENI)

Cloud disk IOPS

Disk bandwidth (Gbit/s)

ecs.ebmg9i.48xlarge

192

768

64/None

20,000,000

6,000,000

64 (primary network interface)/16 (secondary network interface)

32

50

50

500,000/800,000

32/None

General-purpose ebmg8a instance family

  • Introduction: These instances are built on the new Alibaba Cloud CIPU architecture and powered by AMD EPYC Genoa processors. They offer stable compute performance, a more powerful I/O engine, and chip-level security hardening.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Workloads that support third-party hypervisors and require hybrid cloud and multi-cloud deployment.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High packet-rate networking scenarios, such as live comment streaming and telecom traffic forwarding.

    • Enterprise applications of all types and sizes.

    • Web and application servers.

    • Game servers.

    • Small to medium-sized database systems, caches, and search clusters.

    • Data analytics and computing.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of 1:4.

    • Processor: AMD EPYC Genoa processors with a turbo frequency up to 3.7 GHz, which deliver stable compute performance.

  • Storage:

  • Network:

The following table lists the instance types and specifications for the ebmg8a instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth

Network PPS

Connections

Multi-queue

Elastic network interfaces

Private IPv4 addresses

IPv6 addresses

Cloud disk IOPS

Cloud disk bandwidth

ecs.ebmg8a.48xlarge

192

768

64/None

18,000,000

3,000,000

64

38

30

30

500,000/None

32/None

ECS Bare Metal Instance family ebmg8y

  • Introduction: These instances are built on the new Alibaba Cloud CIPU architecture. They provide stable computing power, a powerful I/O engine, dedicated hardware resources, and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers (including but not limited to Docker, Clear Container, and Pouch).

    • Video on demand and live streaming.

    • Enterprise applications of all types and sizes.

    • Web and application servers.

    • Data analytics and computing.

    • High-performance scientific and engineering applications.

  • Compute:

    • Processor: Features Alibaba Cloud's proprietary Yitian 710 ARM processors with a base frequency of at least 2.75 GHz. Hyper-threading is not supported, ensuring stable computing performance.

  • Storage:

  • Network:

This table lists the instance types and specifications of the ebmg8y instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

Queues

ENIs

Private IPv4s/ENI

IPv6 addresses/ENI

Disk IOPS

Disk bandwidth (Gbit/s)

ecs.ebmg8y.32xlarge

128

512

64/None

20,000,000

3,000,000

64 (primary ENI)/32 (secondary ENI)

38

30

30

500,000/None

32/None

General-purpose elastic bare metal instance family ebmg8i

  • Description: Built on Alibaba Cloud's new CIPU architecture, this instance family delivers stable computing power, an enhanced I/O engine, dedicated hardware resources, and physical isolation.

  • Scenarios:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Compatibility with third-party hypervisors for hybrid and multi-cloud deployments.

    • Containers, such as Docker, Clear Containers, and Pouch.

    • High packet-rate networking scenarios, such as live comment streaming and telecommunication service forwarding.

    • Enterprise-level applications of various types and sizes.

    • Websites and application servers.

    • Game servers.

    • Small and medium-sized database systems, caches, and search clusters.

    • Data analytics and computing.

    • High-performance scientific and engineering applications.

  • Compute:

    • Processor: Features Intel® Xeon® Emerald Rapids or Intel® Xeon® Sapphire Rapids processors with a base frequency of at least 2.7 GHz and an all-core turbo frequency of 3.2 GHz.

      Note

      When you purchase an instance, one of the two processor types is randomly assigned. You cannot manually select a processor type.

    • These instances support hyper-threading. Hyper-threading is enabled by default. For more information, see Change CPU options.

    • For information about operating system compatibility, see Intel instance and OS compatibility.

  • Storage:

  • Network:

The following table lists the instance types and specifications for the ebmg8i instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

Multi-queue

ENIs

Private IPv4s per ENI

IPv6 addresses per ENI

Cloud disk IOPS

Cloud disk throughput (Gbit/s)

ecs.ebmg8i.48xlarge

192

1024

100/None

30,000,000

4,000,000

64 (primary ENI)/16 (secondary ENI)

72

30

30

1,000,000/None

48/None

General-purpose Elastic Bare Metal Server instance family ebmg7

  • Instance family overview:

    • Built on the third-generation X-Dragon architecture, these instances use chip-level acceleration and fast-path technology to deliver stable, predictable, and exceptionally high compute, storage, and network performance.

    • These instances provide dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses

    • Compatibility with third-party hypervisors for hybrid and multi-cloud deployments

    • Containers, such as Docker, Clear Container, and Pouch

    • High packet-rate networking scenarios, such as live comment streaming and telecom traffic forwarding

    • Enterprise applications of all types and sizes

    • Website and application servers

    • Game servers

    • Small to medium-sized database systems, caches, and search clusters

    • Data analytics and computing

    • High-performance scientific and engineering applications

  • Compute:

    • vCPU-to-memory ratio of 1:4

    • Processor: Intel® Xeon® Platinum 8369B (Ice Lake) processor with a base frequency of 2.9 GHz and an all-core turbo frequency of 3.5 GHz

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Delivers ultra-high network performance of up to 24 million PPS (packets per second).

The following table lists the instance types and specifications for the ebmg7 instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

ENIs

Private IPv4 addresses

IPv6 addresses

Disk IOPS

Disk bandwidth (Gbit/s)

ecs.ebmg7.32xlarge

128

512

64

24,000,000

2,400,000

32

20

20

600,000

32

General-purpose ECS bare metal instance family ebmg7a

  • Introduction

    • Built on the third-generation X-Dragon architecture, these instances leverage chip-level acceleration and fast-path technology to provide stable, predictable, and high compute, storage, and network performance.

    • These instances provide dedicated hardware resources and physical isolation.

  • Use cases

    • Workloads that require direct access to physical resources or have hardware-based licensing requirements.

    • Compatible with third-party hypervisors for hybrid cloud and multi-cloud deployment.

    • Containers, such as Docker, Clear Container, and Pouch.

    • Compute clusters and memory-intensive data processing.

    • Video encoding/decoding and rendering.

    • Data analytics and computing.

  • Compute

    • vCPU-to-memory ratio of 1:4.

    • These instances use AMD EPYC™ MILAN processors that feature a 2.55 GHz base frequency and a single-core turbo frequency of up to 3.5 GHz to ensure stable computing performance.

  • Storage

  • Network

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances deliver network performance of up to 24 million PPS.

The following table lists the instance types and specifications for the ebmg7a instance family.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Network PPS

Connections

Multi-queue

ENIs

Private IPv4s per ENI

IPv6s per ENI

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.ebmg7a.64xlarge

256

1024

64

24 million

4 million

32

31

15

1

600,000

32

Note
  • ECS instances of this instance type require an image that supports the UEFI boot mode. If you use a custom image, ensure its boot mode is set to UEFI. For instructions on how to set the boot mode for a custom image, see Instance Boot Mode.

  • Do not use Ubuntu 18 or Debian 9 images to create instances of this instance type. The kernels of these operating systems do not support AMD EPYCTM MILAN processors, and the instances will fail to start.

General-purpose ECS bare metal instance family ebmg6a

This instance family is in invitational preview. To use it, submit a ticket.

  • Overview:

    • Built on the third-generation X-Dragon architecture, these instances use chip-level acceleration and fast-path technology to deliver stable, predictable, and ultra-high compute, storage, and network performance.

    • These instances provide dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Containers, and Pouch.

    • Video encoding, decoding, and rendering.

    • Compute clusters and memory-intensive data processing.

    • Data analytics and compute.

  • Compute:

    • The vCPU-to-memory ratio is 1:4.

    • Processor: AMD EPYC ROME processors with a 2.6 GHz base frequency and a 3.3 GHz turbo frequency, delivering consistent compute performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances deliver ultra-high network performance of up to 24 million PPS.

The following table lists the instance types and specifications for the ebmg6a instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Queues

ENIs

Private IPv4 addresses

IPv6 addresses

Cloud disk IOPS

Cloud disk bandwidth (Gbit/s)

ecs.ebmg6a.64xlarge

256

1024

64

24,000,000

32

31

10

1

600,000

32

Note

ECS instances of this instance family must be created from images that use the UEFI boot mode. If you use a custom image, ensure that it is configured to use UEFI. For instructions, see instance boot mode.

Ebmg6e: General-purpose (balanced-enhanced) elastic bare metal instance family

The ebmg6e instance family provides the following features:

  • Instance family overview:

    • Built on the third-generation X-Dragon architecture, these instances use chip-level acceleration and fast-path technology to deliver stable, predictable, and exceptionally high compute, storage, and network performance.

    • These instances provide dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or have hardware-based licensing requirements.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High network packet rate scenarios, such as live comment streaming and telecom service forwarding.

    • Enterprise applications of all types and sizes.

    • Web and application servers.

    • Game servers.

    • Small to medium-sized database systems, caches, and search clusters.

    • Data analytics and computing.

    • Compute clusters and memory-intensive data processing.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of approximately 1:4.

    • Processor: Intel® Xeon® Platinum 8269CY (Cascade Lake) with a base frequency of 2.5 GHz and an all-core turbo frequency of 3.2 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Network performance of up to 24 million PPS.

The following table lists the instance types and specifications for the ebmg6e instance family.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Network PPS

Connections

ENIs

Private IPv4s per ENI

IPv6s per ENI

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.ebmg6e.26xlarge

104

384

32

24,000,000

1,800,000

32

10

1

480,000

16

ebmg6 general-purpose bare metal instance family

The ebmg6 instance family offers the following features:

  • Instance family description: Offers dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • Video encoding, decoding, and rendering.

    • Resource-intensive database applications for medium-to-large enterprises.

    • Compute clusters and memory-intensive data processing.

    • Data analytics.

  • Compute:

    • vCPU-to-memory ratio of approximately 1:4.

    • Processor: Intel® Xeon® Platinum 8269CY (Cascade Lake) with a base frequency of 2.5 GHz and an all-core turbo frequency of 3.2 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • High network performance with up to 6 million PPS.

This table lists the instance types and specifications for the ebmg6 instance family.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Network PPS

Connections

ENIs

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline cloud disk IOPS

Baseline cloud disk bandwidth (Gbit/s)

ecs.ebmg6.26xlarge

104

384

32

6,000,000

1,800,000

32

20

1

200,000

16

Note

By default, CloudMonitor does not provide basic CPU metrics for EBM instances. To collect them, you must install the CloudMonitor agent. For instructions, see Install the CloudMonitor agent.

Memory-optimized elastic bare metal server instance family ebmr9ae

  • Introduction: These instances are built on Alibaba Cloud's new CIPU architecture and powered by the latest AMD EPYC Turin processors. They deliver consistent compute performance, a powerful I/O engine, and chip-level security hardening.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High-performance databases and in-memory databases.

    • Data analytics and mining, and distributed in-memory caches.

    • Hadoop, Spark clusters, and other enterprise applications with high memory requirements.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of 1:8.

    • Processor: AMD EPYC Turin processors with a boost frequency of up to 3.7 GHz. These instances use physical cores to ensure consistent compute performance.

  • Storage:

  • Network:

The following table lists the instance types and specifications for the ebmr9ae instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

Multi-queue

Elastic network interfaces

Private IPv4 addresses

IPv6 addresses

Baseline IOPS

Cloud disk bandwidth (Gbit/s)

ecs.ebmr9ae.48xlarge

192

1536

100/None

30,000,000

6,000,000

64

38

50

50

600,000

50/None

Memory-optimized instance family ebmr9i

  • Instance family description: Built on Alibaba Cloud's new CIPU architecture and powered by Intel® Xeon® 6 processors with P-cores, this instance family delivers consistent compute performance, an enhanced I/O engine, and chip-level security hardening.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High-performance databases and in-memory databases.

    • Data analytics and mining, and distributed in-memory caches.

    • Hadoop and Spark clusters, and other enterprise applications with high memory requirements.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of 1:8.

    • Processor: Intel® Xeon® Granite Rapids processors with a 3.2 GHz base frequency and a 3.6 GHz all-core turbo frequency.

  • Storage:

    • These are I/O optimized instances.

    • These instances support the NVMe protocol. For more information, see Overview of the NVMe protocol.

    • Supported cloud disk types: ESSD Cloud Disks, ESSD AutoPL Cloud Disks, and ESSD Zone-Redundant Cloud Disks. For more information about block storage, see Block Storage Overview.

  • Network:

The following table lists the instance types and specifications for the ebmr9i instance family.

Instance type

vCPU

Memory (GiB)

Baseline/burst network bandwidth (Gbit/s)

Network PPS

Connections

Multi-queue

Elastic network interfaces

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Cloud disk IOPS (baseline/burst)

Cloud disk bandwidth (baseline/burst) (Gbit/s)

ecs.ebmr9i.48xlarge

192

1536

64/None

20,000,000

6,000,000

64 (primary network interface)/16 (secondary network interface)

32

50

50

500,000/800,000

32/None

General-purpose ebmr8a instance family

  • Instance family overview: Built on Alibaba Cloud's next-generation CIPU architecture and powered by AMD EPYC Genoa processors, this instance family delivers consistent compute performance, an enhanced I/O engine, and chip-level security hardening.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Workloads that are compatible with third-party hypervisors for hybrid cloud and multi-cloud deployment.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High-performance databases and in-memory databases.

    • Data analytics and mining, and distributed in-memory caches.

    • Hadoop and Spark clusters, and other enterprise applications with high memory requirements.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of 1:8.

    • Processor: Features AMD EPYC Genoa processors with a boost frequency up to 3.7 GHz for consistent compute performance.

  • Storage:

  • Network:

The ebmr8a instance family includes the following instance types:

Instance type

vCPU

Memory (GiB)

Bandwidth (baseline/burst) (Gbit/s)

Packets per second (PPS)

Connections

Multi-queue

ENIs

Private IPv4s per ENI

IPv6s per ENI

Disk IOPS (baseline/burst)

Disk bandwidth (baseline/burst) (Gbit/s)

ecs.ebmr8a.48xlarge

192

1536

64/None

18,000,000

3,000,000

64

38

30

30

500,000/None

32/None

Memory-enhanced instance family ebmr8y

Note

To use ebmr8y, submit a ticket.

  • Introduction: This instance family uses Alibaba Cloud's next-generation CIPU architecture, providing consistent computing performance, an enhanced I/O engine, dedicated hardware resources, and physical isolation.

  • Scenarios:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Workloads that are compatible with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Containers, and Pouch.

    • Video on demand and live streaming.

    • Enterprise applications of all types and sizes.

    • Websites and application servers.

    • Data analytics and computation.

    • High-performance scientific and engineering applications.

  • Compute:

    • Processor: Features Alibaba Cloud's proprietary Yitian 710 Arm architecture processors with a base frequency of at least 2.75 GHz. These instances do not use hyper-threading, which ensures stable computing performance.

  • Storage:

  • Network:

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

Multi-queue

ENIs

Private IPv4 per ENI

IPv6 per ENI

Cloud disk IOPS

Disk bandwidth (Gbit/s)

ecs.ebmr8y.32xlarge

128

1024

64/None

20,000,000

3,000,000

64 (primary)/32 (secondary)

38

30

30

500,000/None

32/None

Memory-enhanced instance family ebmr8y

Note

To use ebmr8y, submit a ticket.

  • Introduction: This instance family uses Alibaba Cloud's next-generation CIPU architecture, providing consistent computing performance, an enhanced I/O engine, dedicated hardware resources, and physical isolation.

  • Scenarios:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Workloads that are compatible with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Containers, and Pouch.

    • Video on demand and live streaming.

    • Enterprise applications of all types and sizes.

    • Websites and application servers.

    • Data analytics and computation.

    • High-performance scientific and engineering applications.

  • Compute:

    • Processor: Features Alibaba Cloud's proprietary Yitian 710 Arm architecture processors with a base frequency of at least 2.75 GHz. These instances do not use hyper-threading, which ensures stable computing performance.

  • Storage:

  • Network:

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

Multi-queue

ENIs

Private IPv4 per ENI

IPv6 per ENI

Cloud disk IOPS

Disk bandwidth (Gbit/s)

ecs.ebmr8y.32xlarge

128

1024

64/None

20,000,000

3,000,000

64 (primary)/32 (secondary)

38

30

30

500,000/None

32/None

Ebmr7 memory-optimized instance family

  • Introduction

    • Built on the third-generation X-Dragon architecture, these instances use chip-level fast-path acceleration to deliver stable, predictable, and exceptionally high performance for compute, storage, and networking.

    • They provide dedicated hardware resources and physical isolation.

  • Use cases

    • Workloads that require direct access to physical resources or have hardware-bound licensing requirements.

    • Support for third-party hypervisors to enable hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High-performance databases and in-memory databases.

    • Data analytics and mining, and distributed in-memory caches.

    • Hadoop and Spark clusters, and other memory-intensive enterprise applications.

    • High-performance scientific and engineering applications.

  • Compute

    • vCPU-to-memory ratio of 1:8.

    • Processor: Intel® Xeon® Platinum 8369B (Ice Lake) with a base frequency of 2.9 GHz and an all-core turbo frequency of 3.5 GHz.

  • Storage

  • Network

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • They deliver network performance of up to 24 million Network PPS.

The following table shows the specifications for the ebmr7 instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

ENIs

Private IPv4 addresses

IPv6 addresses

Disk IOPS

Disk bandwidth (Gbit/s)

ecs.ebmr7.32xlarge

128

1024

64

24,000,000

2,400,000

32

20

20

600,000

32

Ebmr7a memory-optimized instance family

  • Description:

    • Built on the third-generation X-Dragon architecture, these instances leverage chip-level acceleration and fast-path technology to deliver stable, predictable, and exceptionally high compute, storage, and network performance.

    • These instances provide dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or involve hardware-locked licensing.

    • They are compatible with third-party hypervisors, enabling hybrid cloud and multi-cloud deployments.

    • Running containers such as Docker, Clear Container, and Pouch.

    • Hosting in-memory databases.

    • Powering data analytics, data mining, and distributed in-memory caching.

    • Running Hadoop, Spark clusters, and other enterprise applications that require large amounts of memory.

  • Compute:

    • vCPU-to-memory ratio of 1:8.

    • Processor: 2.55 GHz AMD EPYCTM MILAN processors with a single-core turbo frequency of up to 3.5 GHz provide stable computing performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Offers exceptional network performance with up to 24 million network PPS.

The following table describes the instance types in the ebmr7a instance family.

Instance type

vCPUs

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Network PPS

Connections

Multi-queue

ENIs

Private IPv4 addresses

IPv6 addresses

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.ebmr7a.64xlarge

256

2,048

64

24 million

4 million

32

31

15

1

600,000

32

Note
  • For ECS instances of this instance type, you must use an image with the UEFI boot mode. If you use a custom image, you must ensure that its boot mode is set to UEFI. For details on setting the boot mode for a custom image, see Instance boot modes.

  • Do not use Ubuntu 18 or Debian 9 images for this instance type. Their kernels lack support for AMD EPYCTM MILAN processors, and instances created with them will fail to start.

Memory-optimized bare metal instance family ebmr6a

This instance family is available by invitation only. To request access, submit a ticket.

  • Introduction:

    • Built on the third-generation X-Dragon architecture, these instances use chip-level acceleration and fast-path technology to provide stable, predictable, and high performance for compute, storage, and networking.

    • These instances provide dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • In-memory databases.

    • Data analytics and mining, and distributed in-memory caches.

    • Hadoop, Spark clusters, and other enterprise applications with high memory requirements.

  • Compute:

    • vCPU-to-memory ratio of 1:8.

    • Processors: AMD EPYCTM ROME processors with a base frequency of 2.6 GHz and a turbo frequency of 3.3 GHz for stable compute performance.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances provide ultra-high network performance of up to 24 million PPS.

The following table lists the instance types and specifications for the ebmr6a instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Packet rate (PPS)

Queue pairs

ENIs

Private IPv4 addresses

IPv6 addresses

Cloud disk IOPS

Cloud disk bandwidth (Gbit/s)

ecs.ebmr6a.64xlarge

256

2048

64

24 million

32

31

10

1

600,000

32

Note

ECS instances in this instance family must use images with the boot mode set to UEFI. If you use a custom image, make sure its boot mode is set to UEFI. To learn how to set the boot mode for custom images, see Instance boot mode.

Memory-optimized ebmr6e instance family

  • Introduction:

    • Built on the third-generation X-Dragon architecture, these instances leverage chip-level acceleration and fast-path technology to deliver stable, predictable, and high compute, storage, and network performance.

    • Provides dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or have hardware-based licensing requirements.

    • Third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Containers, and Pouch.

    • Scenarios with high network packet rates, such as video streaming and telecom service forwarding.

    • High-performance databases and in-memory databases.

    • Data analytics and mining, and distributed in-memory caches.

    • Hadoop, Spark clusters, and other enterprise applications with high memory requirements.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of approximately 1:8.

    • Processor: Intel® Xeon® Platinum 8269CY (Cascade Lake) with a base frequency of 2.5 GHz and an all-core turbo frequency of 3.2 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Delivers network performance of up to 24 million PPS.

This table lists the specifications for the ebmr6e instance family.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Network PPS

Connections

Elastic network interfaces

Private IPv4 addresses

IPv6 addresses

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.ebmr6e.26xlarge

104

768

32

24,000,000

1,800,000

32

10

1

480,000

16

Ebmr6 instance family: Memory-optimized elastic bare metal server

  • Introduction: Offers dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licensing.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployment.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High-performance databases and in-memory databases.

    • Data analytics and mining, and distributed in-memory caches.

    • Hadoop, Spark clusters, and other enterprise applications with high memory requirements.

  • Compute:

    • vCPU-to-memory ratio of approximately 1:8.

    • Processors: Intel® Xeon® Platinum 8269CY (Cascade Lake) with a 2.5 GHz base frequency and a 3.2 GHz all-core turbo frequency.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • High network performance: 6,000,000 network PPS.

This table lists the specifications of the ebmr6 instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

ENIs

Private IPv4 addresses

IPv6 addresses

Cloud disk IOPS

Cloud disk bandwidth (Gbit/s)

ecs.ebmr6.26xlarge

104

768

32

6,000,000

1,800,000

32

20

1

200,000

16

Ebmre6p: Persistent memory-optimized instance family

To use ebmre6p, submit a ticket.

  • Description: Provides dedicated hardware resources and physical isolation.

  • Use cases:

    • In-memory databases, such as Redis.

    • High-performance databases, such as SAP HANA.

    • Other memory-intensive applications, such as AI applications and intelligent search applications.

  • Compute:

    • Powered by Intel® Optane persistent memory, these instances are optimized end-to-end for Redis workloads and deliver a high price-performance ratio.

    • These instances support up to 1,920 GiB of memory (384 GiB of DRAM memory and 1,536 GiB of Intel® Optanepersistent memory) with a vCPU-to-memory ratio of nearly 1:20, making them ideal for memory-intensive applications.

    • Processor: 2.5 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) processor with an all-core turbo frequency of 3.2 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances deliver high network performance of up to 6 million packets per second (PPS).

The following table lists the instance types and specifications for the ebmre6p instance family.

Instance type

vCPU

Memory (GiB)

Persistent memory (GiB)

Baseline network bandwidth (Gbit/s)

PPS

ENIs

Private IPv4s/ENI

IPv6s/ENI

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.ebmre6p.26xlarge

104

384

1536

32

6,000,000

31

10

1

200,000

16

Ebmre6-6t memory-enhanced ECS bare metal instance family

To use ebmre6-6t, submit a ticket to request access.

  • Description: Provides dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • High-performance databases and in-memory databases (such as SAP HANA).

    • Memory-intensive applications.

    • Big data processing engines (such as Apache Spark and Presto).

  • Compute:

    • Processor-to-memory ratio of approximately 1:30.

    • Processors: Intel® Xeon® Platinum 8269 (Cascade Lake) with a base frequency of 2.5 GHz and an all-core turbo frequency of 3.2 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Delivers a packet forwarding rate of 6 million PPS.

The following table lists the instance types in the ebmre6-6t instance family.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Network PPS

Connections

ENIs

IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline IOPS

Baseline bandwidth (Gbit/s)

ecs.ebmre6-6t.52xlarge

208

6144

32

6 million

1.8 million

32

10

1

200,000

16

High-frequency general-purpose bare metal instance family ebmhfg7

  • Instance family overview:

    • Built on the third-generation X-Dragon architecture, these instances use chip-level acceleration and fast-path technology to provide stable, predictable, and extremely high performance for compute, storage, and networking.

    • These instances provide dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • High packet-rate networking scenarios, such as live comment streaming and telecom traffic forwarding.

    • Enterprise applications of all types and sizes.

    • Game servers.

    • Small to medium-sized database systems, caches, and search clusters.

    • High-performance scientific computing.

    • Video encoding applications.

  • Compute:

    • vCPU-to-memory ratio of 1:4.

    • Processor: 3rd Generation Intel® Xeon® Scalable processors (Cooper Lake architecture) with a 3.3 GHz base frequency and a 3.8 GHz all-core turbo frequency.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Delivers ultra-high network performance of up to 24 million network PPS.

The following table lists the instance types and specifications for the ebmhfg7 instance family.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Network PPS

Multi-queue

ENIs

Private IPv4 addresses

IPv6 addresses

Baseline IOPS

Baseline disk bandwidth (Gbit/s)

ecs.ebmhfg7.48xlarge

192

768

64

24 million

32

31

10

1

600,000

32

High-frequency compute-optimized instance family ebmhfc7

  • Instance family overview:

    • Powered by the third-generation X-Dragon Architecture, these instances use chip-level fast-path acceleration to deliver stable, predictable, and ultra-high compute, storage, and network performance.

    • They provide dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads requiring direct access to physical resources or with hardware-bound licensing.

    • High-packet-rate networking, such as live comment streaming and telecom traffic forwarding.

    • High-performance front-end server clusters.

    • Massively multiplayer online (MMO) game front-ends.

    • Data analysis, batch computing, and video encoding.

    • High-performance scientific and engineering applications.

  • Compute:

    • vCPU-to-memory ratio of 1:2.

    • Processor: Third-generation Intel® Xeon® Scalable processors (Cooper Lake) with a base frequency of 3.3 GHz and an all-core turbo frequency of 3.8 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Achieves network performance of up to 24 million PPS.

The following table lists the instance types and specifications for the ebmhfc7 instance family.

Instance type

vCPUs

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Network PPS

Multi-queue

ENIs

Private IPv4s per ENI

IPv6s per ENI

Baseline cloud disk IOPS

Baseline cloud disk bandwidth (Gbit/s)

ecs.ebmhfc7.48xlarge

192

384

64

24,000,000

32

31

10

1

600,000

32

High-frequency memory-optimized instance family ebmhfr7

  • Instance family overview:

    • Built on the third-generation X-Dragon architecture, these instances use chip-level acceleration and fast-path technology to deliver stable, predictable, and ultra-high compute, storage, and network performance.

    • Provides dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • High packet-rate networking scenarios, such as live comment streaming and telecom traffic forwarding.

    • High-performance databases and in-memory databases.

    • Data analytics and mining, and distributed in-memory caches.

    • Hadoop and Spark clusters, and other enterprise applications with high memory requirements.

  • Compute:

    • vCPU-to-memory ratio of 1:8.

    • Processor: third-generation Intel® Xeon® Scalable processors (Cooper Lake architecture) with a base frequency of at least 3.3 GHz and an all-core turbo frequency of 3.8 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Delivers ultra-high network performance of up to 24 million network PPS.

This table lists the instance types and specifications of the ebmhfr7 instance family.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Network PPS

Multi-queue

ENIs

Private IPv4 addresses

IPv6 addresses

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.ebmhfr7.48xlarge

192

1536

64

24,000,000

32

31

10

1

600,000

32

Ebmhfg6: a high-frequency, general-purpose bare metal instance family

  • Introduction: Provides dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or have hardware-based licensing requirements.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • Resource-intensive database applications for medium to large enterprises.

    • Video encoding/decoding and rendering.

  • Compute:

    • vCPU-to-memory ratio of 1:4.8.

    • Processor: Intel® Xeon® Platinum 8269CY (Cascade Lake) with a 3.1 GHz base frequency and a 3.5 GHz all-core turbo frequency.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Provides up to 6 million network PPS.

The following table lists the instance types and specifications for the ebmhfg6 instance family.

Instance type

vCPU

Memory (GiB)

Baseline bandwidth (Gbit/s)

Network PPS

Connections

ENIs

Private IPv4s/ENI

IPv6s/ENI

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.ebmhfg6.20xlarge

80

384

32

6,000,000

1,800,000

32

20

1

200,000

16

High-frequency compute-optimized elastic bare metal instance family ebmhfc6

  • Instance family overview: Dedicated hardware resources with physical isolation.

  • Use cases:

    • High packet-rate networking scenarios, such as live comment streaming and telecom traffic forwarding.

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • Video encoding/decoding and rendering.

  • Compute:

    • vCPU-to-memory ratio of 1:2.4

    • Intel® Xeon® Platinum 8269CY (Cascade Lake) processor with a 3.1 GHz base frequency and a 3.5 GHz all-core turbo frequency

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Packet forwarding rate of up to 6 million PPS.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Network PPS

Connections

ENIs

Private IPv4 addresses

IPv6 addresses

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.ebmhfc6.20xlarge

80

192

32

6 million

1.8 million

32

20

1

200,000

16

High-frequency memory-optimized ebmhfr6 instance family

  • Instance family overview: Offers dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or have hardware-bound licensing requirements.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High-performance databases and in-memory databases.

    • Data analytics and mining, and distributed in-memory caching.

    • Hadoop, Spark clusters, and other enterprise applications with large memory requirements.

  • Compute:

    • vCPU-to-memory ratio of 1:9.6.

    • Processor: 3.1 GHz Intel® Xeon® Platinum 8269CY (Cascade Lake) with an all-core turbo frequency of 3.5 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Packet processing rate of up to 6 million PPS.

The following table lists the instance types and specifications for the ebmhfr6 instance family.

Instance type

vCPUs

Memory (GiB)

Baseline network bandwidth (Gbps)

Network PPS

Connections

Elastic network interfaces

Private IPv4 addresses per ENI

IPv6 addresses per ENI

Baseline cloud disk IOPS

Baseline cloud disk bandwidth (Gbps)

ecs.ebmhfr6.20xlarge

80

768

32

6 million

1.8 million

32

20

1

200,000

16

HPC and SCC instance families

HPC-optimized instance family hpc9a

  • About hpc9a: The hpc9a family targets HPC workloads that require large memory capacity, such as chip design. It offers an ultra-high CPU-to-memory ratio of up to 1:24. Built on Alibaba Cloud's latest CIPU architecture and powered by AMD's newest EPYC Turin processors, hpc9a uses physical CPU cores to deliver stable computing power, stronger I/O engines, and chip-level security hardening.

  • Scenarios: Chip design and other HPC workloads.

  • Compute:

    • CPU-to-memory ratio (physical cores to memory) is 1:24. Maximum memory is 3072 GiB.

    • Processor: Fifth-generation AMD EPYC Turin processor. Turbo frequency up to 5.0 GHz. Uses physical cores for stable compute performance.

    • Hyper-threading is not supported.

    • Operating system compatibility: Supports only verified and performance-optimized operating systems, including Alibaba Cloud Linux 3 and 4.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Delivers an ultra-high packet forwarding rate.

    • These instances support eRDMA. For information about how to use eRDMA, see Enable on enterprise-level instance.

    • These instances support jumbo frames. For more information, see Jumbo frames.

hpc9a includes the following instance types and metrics:

Instance type

Physical cores

Memory (GiB)

Baseline network bandwidth (Gbit/s)

eRDMA network bandwidth (Gbit/s)

ecs.hpc9a.32xlarge

128

3072

200

200

Note

This instance family supports Rocky Linux 8 and Rocky Linux 9 images.

HPC-optimized instance family hpc8i

The hpc8i family is in invitational preview. To use it, submit a ticket.

  • About hpc8i: The hpc8i family is optimized for compute-intensive applications such as implicit finite element analysis, molecular dynamics, and computational chemistry. It uses the latest Intel® Xeon® Emerald Rapids processors. All-core turbo frequency is 3.6 GHz. It supports Intel's rich software ecosystem, including Intel Math Kernel Library and Advanced Vector Extensions 512 (AVX-512).

  • Scenarios:

    • Computational Fluid Dynamics (CFD) and Finite Element Analysis (FEA) in industrial simulation.

    • EDA simulation.

    • Geological exploration.

    • Weather forecasting.

    • Molecular dynamics simulation.

    • Other HPC workloads.

  • Compute:

    • CPU-to-memory ratio is 1:8.

    • Processor: Intel® Xeon® Emerald Rapids processor. Base clock speed is at least 2.8 GHz. All-core turbo frequency is 3.6 GHz. Delivers stable compute performance.

    • Hyper-threading is not supported.

    • Operating system compatibility: Supports only verified and performance-optimized operating systems, including Alibaba Cloud Linux 2.1903 LTS 64-bit and Alibaba Cloud Linux 3.2104 LTS 64-bit.

  • Storage:

  • Network:

The hpc8i family includes the following instance types and metrics.

Instance type

Physical cores

Memory (GiB)

Baseline network bandwidth (Gbit/s)

eRDMA network bandwidth (Gbit/s)

ecs.hpc8i.32xlarge

64

512

100

100

Note

This instance family supports Rocky Linux 8 and Rocky Linux 9 images.

HPC-optimized instance family hpc8ae

  • About hpc8ae: The hpc8ae family targets tightly coupled HPC workloads that are both compute- and network-intensive, such as industrial simulation, Electronic Design Automation (EDA) simulation, geological exploration, weather forecasting, and molecular dynamics simulation. It delivers the latest fourth-generation AMD EPYC™ (Genoa) processors with up to 3.75 GHz turbo frequency, 64 Gbps eRDMA inter-node network bandwidth, and enhanced memory bandwidth.

  • Scenarios:

    • Computational Fluid Dynamics (CFD) and Finite Element Analysis (FEA) in industrial simulation.

    • EDA simulation.

    • Geological exploration.

    • Weather forecasting.

    • Molecular dynamics simulation.

    • Other HPC workloads.

  • Compute:

    • CPU-to-memory ratio is 1:4.

    • Processor: AMD EPYC Genoa processor with 3.4 GHz base clock speed. Single-core turbo frequency up to 3.75 GHz. Delivers stable compute performance.

    • Hyper-threading is not supported.

    • Operating system compatibility: Supports only verified and performance-optimized operating systems, including Alibaba Cloud Linux 2.1903 LTS 64-bit and Alibaba Cloud Linux 3.2104 LTS 64-bit.

  • Storage:

  • Network:

The hpc8ae family includes the following instance types and metrics.

Instance type

Physical cores

Memory (GiB)

Baseline network bandwidth (Gbit/s)

eRDMA network bandwidth (Gbit/s)

ecs.hpc8ae.32xlarge

64

256

64

64

Note

This instance family supports Rocky Linux 8 and Rocky Linux 9 images.

HPC-optimized instance family hpc7ip

  • About hpc7ip: The hpc7ip family targets HPC workloads that require large memory capacity, such as chip design. Built on the third-generation SHENLONG architecture, it delivers stable and predictable high performance. Chip-level fast-path acceleration boosts storage and network performance and improves compute stability by orders of magnitude. It offers an ultra-high CPU-to-memory ratio of up to 1:32 and uses Intel Optane persistent memory to significantly reduce the cost per GiB of memory for memory-intensive applications.

  • Scenarios: Chip design and other HPC workloads.

  • Compute:

    • CPU-to-memory ratio (DRAM + persistent memory) is approximately 1:32.

    • Processor: Third-generation Intel® Xeon® Scalable processor (Ice Lake). Base clock speed is 2.9 GHz. All-core turbo frequency is 3.5 GHz. Delivers stable compute performance.

    • Maximum memory is 2560 GiB (512 GiB DRAM + 2048 GiB Intel® Optane persistent memory).

    • Hyper-threading is not supported.

    • Operating system compatibility: Supports only verified and performance-optimized operating systems, including Alibaba Cloud Linux 2.1903 LTS 64-bit and Alibaba Cloud Linux 3.2104 LTS 64-bit.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances provide an ultra-high packet forwarding rate.

The hpc7ip family includes the following instance types and metrics.

Instance type

Physical cores

Memory (GiB)

Persistent memory (GiB)

Baseline network bandwidth (Gbit/s)

ecs.hpc7ip.32xlarge

64

512

2048

64

Note

This instance family supports Rocky Linux 8 and Rocky Linux 9 images.

HPC-optimized instance family hpc6id

  • About hpc6id: The hpc6id family targets HPC workloads that require large memory capacity and local data access, such as chip design. Built on the third-generation SHENLONG architecture, it delivers stable and predictable high performance. Chip-level fast-path acceleration boosts storage and network performance and improves compute stability by orders of magnitude. It provides large memory capacity and two 3.8 TB local data disks to reduce costs for memory- and data-bound applications.

  • Scenarios: Chip design, seismic reservoir and structural simulation, and other HPC workloads.

  • Compute:

    • CPU-to-memory ratio is approximately 1:38.

    • Processor: Intel® Xeon® Scalable processor (Cascade Lake). Base clock speed is 3.1 GHz. All-core turbo frequency is 3.5 GHz. Delivers stable compute performance.

    • Hyper-threading is not supported.

    • Operating system compatibility: Supports only verified and performance-optimized operating systems, including Alibaba Cloud Linux 2.1903 LTS 64-bit and Alibaba Cloud Linux 3.2104 LTS 64-bit.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • These instances provide an ultra-high packet forwarding rate.

The hpc6id family includes the following instance types and metrics.

Instance type

Physical cores

Memory (GiB)

Local storage (GB)

Baseline network bandwidth (Gbit/s)

ecs.hpc6id.20xlarge

40

1536

2 × 3840

32

Note

This instance family supports Rocky Linux 8 and Rocky Linux 9 images.

sccg7, general-purpose SCC instance family

  • Introduction: This instance family provides all features of ECS Bare Metal Instance. For more information, see Overview of ECS Bare Metal Instance families.

  • Supported scenarios:

    • Large-scale machine learning training

    • Large-scale high-performance scientific computing and simulations

    • Large-scale data analytics, batch processing, and video encoding

  • Compute:

    • Offers a CPU-to-memory ratio of 1:4.

    • Uses 2.9 GHz Intel® Xeon® Platinum 8369 (Ice Lake) processors that deliver an all-core turbo frequency of 3.5 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Supports both RoCE networks and VPCs. RoCE networks are dedicated to RDMA communication.

sccg7 instance types

Instance type

vCPUs

Physical cores

Memory (GiB)

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (pps)

RoCE network bandwidth (Gbit/s)

ENIs

ecs.sccg7.32xlarge

128

64

512.0

100

24,000,000

200

32

sccc7, compute-optimized SCC instance family

  • Introduction: This instance family provides all features of ECS Bare Metal Instance. For more information, see Overview of ECS Bare Metal Instance families.

  • Supported scenarios:

    • Large-scale machine learning training

    • Large-scale high-performance scientific computing and simulations

    • Large-scale data analytics, batch processing, and video encoding

  • Compute:

    • Offers a CPU-to-memory ratio of 1:2.

    • Uses 2.9 GHz Intel® Xeon® Platinum 8369 (Ice Lake) processors that deliver an all-core turbo frequency of 3.5 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Supports both RoCE networks and VPCs. RoCE networks are dedicated to RDMA communication.

sccc7 instance types

Instance type

vCPUs

Physical cores

Memory (GiB)

Network baseline bandwidth (Gbit/s)

Packet forwarding rate (pps)

RoCE network bandwidth (Gbit/s)

ENIs

ecs.sccc7.32xlarge

128

64

256.0

100

24,000,000

200

32

Ebmc5s instance family

  • Instance family description: Dedicated hardware resources and physical isolation.

  • Use cases:

    • High packet-rate networking workloads, such as live comment overlays and telecom service forwarding.

    • Workloads that require direct access to physical resources or hardware-dependent licensing.

    • Running third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • Video encoding/decoding and rendering.

  • Compute:

    • vCPU-to-memory ratio of 1:2.

    • Processor: Intel® Xeon® Platinum 8163 (Skylake) with a base frequency of 2.5 GHz and an all-core turbo of 2.7 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Achieves up to 4,500,000 network PPS.

The following table lists the instance types and specifications for the ebmc5s instance family.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Network PPS

Connections

Elastic network interfaces

Private IPv4 addresses

IPv6 addresses

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.ebmc5s.24xlarge

96

192

32

4,500,000

1,800,000

32

10

1

200,000

16

Ebmg5s instance family

  • Instance family overview: Offers dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Workloads that are compatible with third-party hypervisors for hybrid cloud and multi-cloud deployment.

    • Containers such as Docker, Clear Container, and Pouch.

    • Resource-intensive database applications for medium to large enterprises.

    • Video encoding.

  • Compute:

    • vCPU-to-memory ratio of 1:4.

    • Processor: Intel® Xeon® Platinum 8163 (Skylake) with a base frequency of 2.5 GHz and an all-core turbo frequency of 2.7 GHz.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For information about IPv6 communication, see IPv6 communication.

    • Packet forwarding rates up to 4.5 million PPS.

The following table lists the instance types and specifications for the ebmg5s instance family.

Instance type

vCPU

Memory (GiB)

Baseline network bandwidth (Gbit/s)

Network PPS

Connections

ENIs

Private IPv4 addresses

IPv6 addresses

Baseline disk IOPS

Baseline disk bandwidth (Gbit/s)

ecs.ebmg5s.24xlarge

96

384

32

4.5 million

1.8 million

32

10

1

200,000

16

Memory-optimized and network-enhanced ebmr5s family

  • Instance family description: Provides dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • High-performance databases and in-memory databases.

    • Data analytics and mining, and distributed in-memory caches.

    • Hadoop and Spark clusters, and other enterprise applications with high memory requirements.

  • Compute:

    • vCPU-to-memory ratio of 1:8.

    • Processor: Intel® Xeon® Platinum 8163 (Skylake) with a base frequency of 2.5 GHz and an all-core turbo frequency of 2.7 GHz.

  • Storage:

  • Network:

    • Supports IPv4 only.

    • Achieves a packet forwarding rate of up to 4.5 million PPS.

This table lists specifications for the ebmr5s instance family.

Instance type

vCPU

Memory (GiB)

Network bandwidth (Gbit/s)

Network PPS

Connections

ENIs

Private IPv4s per ENI

Disk IOPS

Disk bandwidth (Gbit/s)

ecs.ebmr5s.24xlarge

96

768

32

4,500,000

1,800,000

32

10

200,000

16

ebmg5: General-purpose bare metal instance family

  • Description: Provides dedicated hardware resources and physical isolation.

  • Use cases:

    • Workloads that require direct access to physical resources or hardware-bound licenses.

    • Compatibility with third-party hypervisors for hybrid cloud and multi-cloud deployments.

    • Containers, such as Docker, Clear Container, and Pouch.

    • Resource-intensive database workloads for medium and large enterprises.

    • Video encoding.

  • Compute:

    • vCPU-to-memory ratio of 1:4.

    • Processor: Intel® Xeon® Platinum 8163 (Skylake) with a base frequency of 2.5 GHz and an all-core turbo frequency of 2.7 GHz.

  • Storage:

    • These are I/O optimized instances.

    • Supported cloud disk types: SSD cloud disks and ultra cloud disks. For more information about cloud disks, see Block storage overview.

  • Network:

    • Supports IPv4 only.

    • High network performance, up to 4 million network PPS.

The following table lists the instance types and specifications of the ebmg5 instance family.

Instance type

vCPU

Memory (GiB)

Baseline bandwidth (Gbit/s)

Network PPS

Elastic network interfaces

IPv4s per ENI

ecs.ebmg5.24xlarge

96

384

10

4.5 million

32

10

Heterogeneous computing instance families

sgn8ia vGPU-accelerated instance family

  • Overview:

    • Powered by the third-generation SHENLONG architecture, sgn8ia instances deliver stable and predictable ultra-high performance. Chip-level acceleration provides a significant boost in storage, network, and compute performance, helping you store data and load models faster.

    • These instances include an NVIDIA GRID Virtual Workstation (vWS) software license that provides certified graphics acceleration for a variety of professional computer-aided design (CAD) applications. They can also serve as cost-effective, lightweight GPU-accelerated instances for small-scale AI reasoning workloads.

  • Use cases:

    • Concurrent AI reasoning workloads that require high-frequency CPUs, ample memory, and powerful GPUs, such as image recognition, speech recognition, and behavior identification.

    • Graphics-intensive applications that require high-performance 3D graphics virtualization, such as remote graphic design and cloud gaming. These instances support RTX features and are paired with high-frequency CPUs.

    • 3D modeling for film and animation production, cloud gaming, and mechanical design. The high-frequency AMD Genoa processors, with clock speeds up to 3.75 GHz, provide superior performance for these tasks.

  • Compute:

    • Equipped with NVIDIA Lovelace architecture GPUs that feature:

      • Large GPU memory and multiple GPU slicing options.

      • Support for common acceleration features such as vGPU, RTX, and TensorRT for a wide range of workloads.

    • Processor: High-frequency AMD Genoa processors with clock speeds from 3.4 GHz to 3.75 GHz, delivering higher compute power for 3D modeling.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For more information, see IPv6 communication.

    • Network performance scales with the instance type, as larger instances offer better performance.

The following table lists the instance types and specifications for the sgn8ia family.

Instance type

vCPUs

Memory (GiB)

GPU memory

Network bandwidth (Gbit/s)

Packet rate (pps)

NIC queues

ENIs

Private IPs per ENI

Maximum cloud disks

Baseline IOPS

Baseline throughput (MB/s)

ecs.sgn8ia-m2.xlarge

4

16

2 GB

2.5

1,000,000

4

4

15/15

9

30,000

244

ecs.sgn8ia-m4.2xlarge

8

32

4 GB

4

1,600,000

8

4

15/15

9

45,000

305

ecs.sgn8ia-m8.4xlarge

16

64

8 GB

7

2,000,000

16

8

30/30

17

60,000

427

ecs.sgn8ia-m16.8xlarge

32

128

16 GB

10

3,000,000

32

8

30/30

33

80,000

610

ecs.sgn8ia-m24.12xlarge

48

192

24 GB

16

4,500,000

48

8

30/30

33

120,000

1000

ecs.sgn8ia-m48.24xlarge

96

384

48 GB

32

9,000,000

64

15

30/30

33

240,000

2000

Note
  • The GPU specifications in the table refer to vGPU slices created using vGPU technology.

  • For sgn8ia instances, the memory and GPU memory are dedicated resources. The vCPUs are shared resources with an oversubscription ratio of approximately 1:1.5. If your workload requires dedicated CPU computing power, use a dedicated instance with passthrough GPUs, such as an instance from the gn7i GPU-accelerated compute-optimized family.

sgn7i-vws, vGPU-accelerated instance family with shared CPUs

  • Overview:

    • Powered by the third-generation SHENLONG architecture, sgn7i-vws instances deliver stable, predictable, and high performance. They use chip-level fast path acceleration to improve storage, network performance, and compute stability, helping you store data and load models faster.

    • Instances in this instance family share CPU and network resources to maximize underlying resource utilization. Memory and GPU memory are dedicated to each instance, ensuring data isolation and consistent performance.

      Note

      If you need dedicated CPU resources, select the vgn7i-vws instance family.

    • These instances include an NVIDIA GRID Virtual Workstation (vWS) software license, which provides certified graphics acceleration drivers for various professional CAD applications. They can also be used as lightweight, cost-effective GPU compute instances for small-scale AI inference workloads.

  • Use cases:

    • Equipped with high-performance CPUs, memory, and GPUs, these instances are suitable for concurrent AI inference tasks such as image recognition, speech recognition, and behavior recognition.

    • Supporting RTX features and high-frequency CPUs, these instances provide high-performance 3D graphics virtualization, ideal for graphics-intensive workloads such as remote graphic design and cloud gaming.

    • Powered by Intel Ice Lake processors, these instances deliver excellent performance for 3D modeling in fields such as film and animation production, cloud gaming, and mechanical design.

  • Compute:

    • Equipped with NVIDIA A10 GPUs.

      • Innovative NVIDIA Ampere architecture.

      • Support for common acceleration features such as vGPU, RTX, and TensorRT for a variety of workloads.

    • Processor: 2.9 GHz Intel® Xeon® Scalable (Ice Lake) processor with an all-core turbo frequency of 3.5 GHz.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For more information about IPv6 communication, see IPv6 communication.

    • Network performance scales with the instance type; larger types offer higher performance.

The sgn7i-vws instance family includes the instance types and specifications listed in the following table:

Instance type

vCPU

Memory (GiB)

GPU

GPU memory

Network bandwidth (Gbit/s)

Packet rate (PPS)

NIC queues

ENIs

Private IPv4 addresses

IPv6 addresses

ecs.sgn7i-vws-m2.xlarge

4

15.5

NVIDIA A10 * 1/12

24 GB * 1/12

1.5/5

500,000

4

2

2

1

ecs.sgn7i-vws-m4.2xlarge

8

31

NVIDIA A10 * 1/6

24 GB * 1/6

2.6/10

1,000,000

4

4

6

1

ecs.sgn7i-vws-m8.4xlarge

16

62

NVIDIA A10 * 1/3

24 GB * 1/3

5/20

2,000,000

8

4

10

1

ecs.sgn7i-vws-m2s.xlarge

4

8

NVIDIA A10 * 1/12

24 GB * 1/12

1.5/5

500,000

4

2

2

1

ecs.sgn7i-vws-m4s.2xlarge

8

16

NVIDIA A10 * 1/6

24 GB * 1/6

2.6/10

1,000,000

4

4

6

1

ecs.sgn7i-vws-m8s.4xlarge

16

32

NVIDIA A10 * 1/3

24 GB * 1/3

5/20

2,000,000

8

4

10

1

Note

The GPU column shows the GPU model and GPU slicing information. GPU slicing divides a physical GPU into multiple slices and allocates one slice to each instance.

For example, in NVIDIA A10 * 1/12, NVIDIA A10 is the GPU model, and 1/12 means that a single physical GPU is divided into 12 slices, with one slice allocated to the instance.

vgn7i-vws, vGPU-accelerated instance family

  • Overview

    • Built on the third-generation Shenlong architecture, these instances deliver stable and predictable high performance. They use chip-level fast path acceleration to significantly improve storage performance, network performance, and compute stability, allowing you to store data and load models faster.

    • These instances include an NVIDIA GRID vWS software license, providing certified graphics acceleration drivers for professional CAD applications and graphic design workloads. They can also be used as cost-effective, lightweight compute instances for small-scale AI inference.

  • Use cases

    • With high-performance CPUs, memory, and GPUs, these instances can handle a high volume of concurrent AI inference tasks, such as image recognition, speech recognition, and behavior recognition.

    • These instances support RTX features and are paired with high-frequency CPUs to provide high-performance 3D graphics virtualization, making them ideal for remote graphic design, cloud gaming, and other demanding graphics processing workloads.

    • Powered by Intel Ice Lake processors, these instances deliver excellent performance for 3D modeling in fields such as film and animation production, cloud gaming, and mechanical design.

  • Compute

    • These instances are equipped with NVIDIA A10 GPUs.

      • They are based on the innovative NVIDIA Ampere architecture.

      • They support common acceleration features such as vGPU, RTX, and TensorRT.

    • Processor: 2.9 GHz Intel® Xeon® Scalable (Ice Lake) processors with an all-core turbo frequency of 3.5 GHz.

  • Storage

  • Network

    • These instances support IPv4 and IPv6. For more information about IPv6 communication, see IPv6 communication.

    • Network performance scales with the instance type. Larger instance types offer better network performance.

The vgn7i-vws instance family includes the following instance types and specifications.

Instance type

vCPUs

Memory (GiB)

GPU

GPU memory

Baseline bandwidth (Gbit/s)

Forwarding rate (pps)

NIC queues

ENIs

Private IPv4 addresses

IPv6 addresses

ecs.vgn7i-vws-m4.xlarge

4

30

NVIDIA A10 * 1/6

24GB * 1/6

3

1,000,000

4

4

10

1

ecs.vgn7i-vws-m8.2xlarge

10

62

NVIDIA A10 * 1/3

24GB * 1/3

5

2,000,000

8

6

10

1

ecs.vgn7i-vws-m12.3xlarge

14

93

NVIDIA A10 * 1/2

24GB * 1/2

8

3,000,000

8

6

15

1

ecs.vgn7i-vws-m24.7xlarge

30

186

NVIDIA A10 * 1

24GB * 1

16

6,000,000

12

8

30

1

Note

The GPU column in the preceding table indicates the GPU model and GPU slicing information. GPU slicing divides a physical GPU into multiple slices, allowing each instance to use one slice. For example:

In NVIDIA A10 * 1/6, NVIDIA A10 is the GPU model, and 1/6 indicates that a physical GPU is divided into six slices and the instance uses one of them.

vgn6i-vws, vGPU-accelerated instance family

Important
  • Alibaba Cloud has upgraded the vgn6i instance family to vgn6i-vws. The new instance family uses the latest NVIDIA GRID driver and includes a complimentary GRID vWS license. To request a free image with the GRID driver pre-installed, submit a ticket.

  • If you need to use a public image or a custom image that does not include the GRID driver, submit a ticket to request the GRID driver package for separate installation. Alibaba Cloud does not charge an additional license fee for the GRID driver.

  • Use cases:

    • Real-time rendering for cloud gaming.

    • Real-time rendering for augmented reality (AR) and virtual reality (VR) applications.

    • AI (DL and ML) inference for elastically deployed internet services.

    • Hands-on training environments for deep learning.

    • Model experimentation environments for deep learning.

  • Compute:

    • Equipped with NVIDIA T4 GPU accelerators.

    • These instances use vGPUs created by GPU slicing.

      • Provides 1/4 or 1/2 the compute capability of a full NVIDIA Tesla T4 GPU.

      • Offers 4 GB or 8 GB of GPU memory.

    • Features a vCPU-to-memory ratio of approximately 1:5.

    • Processor: 2.5 GHz Intel ® Xeon ® Platinum 8163 (Skylake).

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For more information, see IPv6 communication.

    • Network performance scales with the instance type.

The following table lists the instance types and specifications for the vgn6i-vws instance family.

Instance type

vCPUs

Memory (GiB)

GPU

GPU memory

Baseline bandwidth (Gbit/s)

Forwarding rate (pps)

NIC queues

ENIs

Private IPv4 addresses

IPv6 addresses

ecs.vgn6i-m4-vws.xlarge

4

23

NVIDIA T4 * 1/4

16 GB * 1/4

2

500,000

4/2

3

10

1

ecs.vgn6i-m8-vws.2xlarge

10

46

NVIDIA T4 * 1/2

16 GB * 1/2

4

800,000

8/2

4

10

1

ecs.vgn6i-m16-vws.5xlarge

20

92

NVIDIA T4 * 1

16 GB * 1

7.5

1,200,000

6

4

10

1

Note

The GPU column in the preceding table indicates the GPU model and GPU slicing information. GPU slicing divides a physical GPU into multiple slices, and each instance uses one slice.

In NVIDIA T4 * 1/4, NVIDIA T4 is the GPU model, and 1/4 indicates that one GPU is divided into four slices, with each instance using one.

gn9gc, GPU-accelerated compute-optimized instance family

Note

gn9gc is in invitational preview. To use gn9gc, submit a ticket.

  • Overview: gn9gc is Alibaba Cloud's 9th-generation cost-effective GPU cloud server instance family. It uses the latest-generation CIPU 2.0 to deliver cloud service capabilities, features high clock speed processors, and is configured with appropriate memory capacity. This instance family provides cost-effective instances for large language model (LLM) generation scenarios and video/image generation scenarios. The GPU can also directly provide graphics processing capabilities to support various rendering workloads.

  • Use cases:

    • LLM inference: The new-generation GPU delivers compute power beyond the 8th generation with significantly improved memory bandwidth. Newly supported FP4 compute comprehensively improves inference performance and cost-effectiveness. Multi-GPU parallel inference efficiency is greatly enhanced.

  • Compute:

    • Uses the latest CIPU 2.0 cloud processor.

      • The 2nd-generation CIPU provides higher cloud processing power with enhanced eRDMA, VPC, and EBS component capabilities. Supports containers (including but not limited to Docker, Clear Container, and Pouch).

    • Uses the new Blackwell architecture professional graphics card:

      • Supports OpenGL professional-grade graphics processing.

      • Supports RTX, TensorRT, and other common acceleration features, with newly upgraded FP4 support and PCIe Gen5 interconnect.

    • Key GPU specifications:

      GPU architecture

      GPU memory

      Computing performance

      Video encoding/decoding

      Inter-GPU interconnect

      Acceleration APIs

      NVIDIA Blackwell

      • Capacity: 72 GB

      • Bandwidth: 1,344 GB/s

      • TF32: 126 TFLOPS

      • FP32: 52 TFLOPS

      • FP16/BF16: 266 TFLOPS

      • FP8/INT8: 530 TFLOPS

      • FP4: 970 TFLOPS

      • RT Core: 196 TFLOPS

      • 3 x Video Encoder

      • 3 x Video Decoder

      • PCIe interface: PCIe Gen5 x16

      • Bandwidth: 128 GB/s, P2P supported

      DX12, OpenGL 4.6, Vulkan 1.3, CUDA 12.8, OpenCL 3.0, DirectCompute

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For more information about IPv6, see IPv6.

    • Ultra-high network performance with up to 30 million PPS (8-GPU instances).

    • Supports ERI (Elastic RDMA Interface) for RDMA direct acceleration over VPC networks, with bandwidth up to 360 Gbit/s. Suitable for autonomous driving, embodied intelligence, computer vision, and traditional model training workloads.

    • Note

      For more information about ERI, see Enable eRDMA on an enterprise-level instance.

The following table describes the instance types in the gn9gc instance family.

Instance type

vCPUs

Memory (GiB)

GPU memory

Baseline/burst bandwidth (Gbit/s)

Packet forwarding rate (pps)

IPv4 addresses per ENI

IPv6 addresses per ENI

NIC queues (primary/secondary)

ENIs

Max data disks

Max disk bandwidth (GB/s)

ecs.gn9gc.4xlarge

16

128

72 GB × 1

16

3.6 million

30

30

8/32

8

1

1

ecs.gn9gc.8xlarge

32

192

72 GB × 1

32

7.5 million

30

30

16/64

8

1

1

ecs.gn9gc-2x.16xlarge

64

384

72 GB × 2

65

15 million

30

30

32/64

15

2

2

ecs.gn9gc-4x.32xlarge

128

768

72 GB × 4

131

30 million

50

50

64/64

15

4

4

ecs.gn9gc-8x.64xlarge

256

1,536

72 GB × 8

204

30 million

50

50

128/64

15

6

6

Note

Images used for gn9gc instances must be in the UEFI boot mode. If you want to use a custom image, make sure that the custom image supports UEFI boot mode and that the boot mode attribute of the image is set to UEFI. For more information, see Set the boot mode of a custom image to UEFI by calling API operations.

gn8v and gn8v-tee, GPU-accelerated compute-optimized instance family

These instance families are available in select regions, including those outside the Chinese mainland. To use them, contact your Alibaba Cloud sales representative.

  • Introduction:

    • gn8v: An 8th-generation GPU-accelerated compute-optimized instance family from Alibaba Cloud for AI model training and inference on ultra-large language models (LLMs). This family provides instance types with one, two, four, or eight GPUs for various application requirements.

    • gn8v-tee: To enhance security for large model training and inference, Alibaba Cloud offers gn8v-tee, an 8th-generation instance family based on gn8v with a confidential computing feature. These instances encrypt data during GPU computation to protect your data.

  • Use cases:

    • Cost-effective for multi-GPU parallel inference on LLMs with more than 70 billion parameters.

    • Each GPU provides 39.5 TFLOPS of FP32 compute power and delivers outstanding performance for traditional AI model training and autonomous driving training workloads.

    • The eight GPUs support NVLink interconnectivity and are suitable for training small- to medium-sized models.

  • Features:

    • High-speed, large-capacity GPU memory: Each GPU is equipped with 96 GB of HBM3 GPU memory and provides up to 4 TB/s of memory bandwidth, significantly accelerating model training and inference.

    • High inter-GPU bandwidth: Multiple GPUs are interconnected with NVLink at 900 GB/s. This enables much higher efficiency for multi-GPU training and inference compared to previous-generation GPU instances.

    • LLM quantization: Supports FP8 compute power, which optimizes performance for large-scale parameter training and inference. This significantly improves training and inference speeds and reduces GPU memory usage.

    • (For gn8v-tee instances only) High security: Supports both CPU confidential computing with Intel® Trust Domain Extensions (TDX) and GPU confidential computing with NVIDIA Confidential Computing (CC). This provides end-to-end confidential computing for the entire model inference pipeline, protecting your inference data and enterprise models during model training and inference.

  • Compute:

    • Powered by the latest CIPU 1.0.

      • Decouples compute from storage, letting you flexibly select the storage resources you need.

      • Provides bare metal capabilities, which support peer-to-peer (P2P) communication between GPU instances, unlike traditional virtualized instances.

    • Powered by 4th-generation Intel® Xeon® Scalable processors with a base frequency of up to 2.8 GHz and an all-core turbo frequency of up to 3.1 GHz.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For more information about IPv6 communication, see IPv6 communication.

    • These instances support jumbo frames. For more information, see Jumbo frames.

    • Delivers ultra-high network performance with a packet forwarding rate of up to 30 million pps (on 8-GPU instances).

    • Supports elastic RDMA interface (ERI).

    • Note

      For information about how to use ERI, see Enable on enterprise-level instance.

  • Security: Supports the trusted computing feature (vTPM). This feature is available on gn8v instances but not on gn8v-tee instances. For more information, see Overview of trusted computing capabilities.

The following table describes the instance types in the gn8v family.

Instance type

vCPUs

Memory (GiB)

GPU memory

Network bandwidth (Gbit/s)

ENIs

Primary ENI queues

IPv4 addresses per ENI

IPv6 addresses per ENI

Max cloud disks

Baseline IOPS

Baseline bandwidth (GB/s)

ecs.gn8v.4xlarge

16

96

96 GB × 1

12

8

16

30

30

17

100,000

0.75

ecs.gn8v.6xlarge

24

128

96 GB × 1

15

8

24

30

30

17

120,000

0.937

ecs.gn8v-2x.8xlarge

32

192

96 GB × 2

20

8

32

30

30

25

200,000

1.25

ecs.gn8v-4x.8xlarge

32

384

96 GB × 4

20

8

32

30

30

25

200,000

1.25

ecs.gn8v-2x.12xlarge

48

256

96 GB × 2

25

8

48

30

30

33

300,000

1.50

ecs.gn8v-8x.16xlarge

64

768

96 GB × 8

32

8

64

30

30

33

360,000

2.5

ecs.gn8v-4x.24xlarge

96

512

96 GB × 4

50

15

64

30

30

49

500,000

3

ecs.gn8v-8x.48xlarge

192

1024

96 GB × 8

100

15

64

50

50

65

1,000,000

6

The following table describes the instance types in the gn8v-tee family.

Instance type

vCPUs

Memory (GiB)

GPU memory

Network bandwidth (Gbit/s)

ENIs

Primary ENI queues

IPv4 addresses per ENI

IPv6 addresses per ENI

Max cloud disks

Baseline IOPS

Baseline bandwidth (GB/s)

ecs.gn8v-tee.4xlarge

16

96

96 GB × 1

12

8

16

30

30

17

100,000

0.75

ecs.gn8v-tee.6xlarge

24

128

96 GB × 1

15

8

24

30

30

17

120,000

0.937

ecs.gn8v-tee-8x.16xlarge

64

768

96 GB × 8

32

8

64

30

30

33

360,000

2.5

ecs.gn8v-tee-8x.48xlarge

192

1024

96 GB × 8

100

15

64

50

50

65

1,000,000

6

Note

The gn8v-tee instance family supports only Alibaba Cloud Linux 3 images. If you use a custom image built on Alibaba Cloud Linux 3 to create an instance, ensure the kernel version is 5.10.134-18 or later.

gn8is, GPU-accelerated compute-optimized instance family

This instance family is available in select regions, including those outside the Chinese mainland. To use this instance family, contact your Alibaba Cloud sales representative.

  • Introduction: gn8is is Alibaba Cloud's eighth-generation GPU-accelerated compute-optimized instance family, designed for the growing demands of AI-generated content (AIGC). Powered by the latest NVIDIA L20 GPUs, this family offers instance types with one, two, four, or eight GPUs, and various CPU-to-GPU ratios to meet diverse application needs.

  • Features:

    • Graphics processing: Powered by 4th-generation Intel® Xeon® Scalable high-frequency processors, these instances provide robust CPU compute power for 3D modeling scenarios, ensuring smoother graphics rendering and design workflows.

    • Inference tasks: Equipped with the new NVIDIA L20 GPU, each with 48 GB of GPU memory, these instances accelerate inference tasks. They support the FP8 floating-point format and can be paired with Container Service for Kubernetes (ACK) to flexibly run inference for various AIGC models. They are especially suitable for inference tasks on large language models (LLMs) with fewer than 70 billion parameters.

  • Use cases:

    • Use GRID drivers with images from Alibaba Cloud Marketplace to enable OpenGL and Direct3D capabilities. This provides workstation-grade graphics processing for workloads such as animation, film and television special effects, and rendering.

    • Use the container management capabilities of Container Service for Kubernetes (ACK) for more efficient and cost-effective AIGC image generation and LLM inference.

    • Other general-purpose AI applications, such as image recognition and speech recognition.

  • Compute:

    • Powered by the new NVIDIA L20 enterprise-grade GPUs.

      • Supports common acceleration features such as TensorRT and the FP8 floating-point format to improve model inference performance.

      • Up to 48 GB of GPU memory per GPU. With multiple GPUs, instances in this family support single-instance inference for models with 70 billion or more parameters.

      • Enhanced graphics processing capabilities. After you install a GRID driver using Cloud Assistant or an image from Alibaba Cloud Marketplace, the graphics processing performance is twice that of 7th-generation platforms.

    • Key parameters of the NVIDIA L20 GPU:

      GPU architecture

      GPU memory

      Compute performance

      Video encoding/decoding

      Inter-GPU connectivity

      NVIDIA Ada Lovelace

      • Capacity: 48 GB

      • Bandwidth: 864 GB/s

      • FP64: N/A

      • FP32: 59.3 TFLOPS

      • FP16/BF16: 119 TFLOPS

      • FP8/INT8: 237 TFLOPS

      • 3 × Video Encoders (+AV1)

      • 3 × Video Decoders

      • 4 × JPEG Decoders

      • PCIe interface: PCIe Gen4 x16

      • Bandwidth: 64 GB/s

    • Processor: Powered by the latest high-frequency Intel® Xeon® processors with an all-core turbo frequency of up to 3.9 GHz to handle complex 3D modeling demands.

  • Storage:

  • Network:

  • Security: These instances support the vTPM feature. For more information, see Overview of trusted computing.

The following table describes the instance types and specifications for the gn8is family.

Instance type

vCPUs

Memory (GiB)

GPU

GPU memory

Network bandwidth (Gbit/s)

ENIs

Primary ENI queues

Private IPv4 addresses

IPv6 addresses

Max cloud disks

Disk IOPS

Disk bandwidth (GB/s)

ecs.gn8is.2xlarge

8

64

L20 × 1

48 GB × 1

8

4

8

15

15

17

60,000

0.75

ecs.gn8is.4xlarge

16

128

L20 × 1

48 GB × 1

16

8

16

30

30

17

120,000

1.25

ecs.gn8is-2x.8xlarge

32

256

L20 × 2

48 GB × 2

32

8

32

30

30

33

250,000

2

ecs.gn8is-4x.16xlarge

64

512

L20 × 4

48 GB × 4

64

8

64

30

30

33

450,000

4

ecs.gn8is-8x.32xlarge

128

1024

L20 × 8

48 GB × 8

100

15

64

50

50

65

900,000

8

gn7e, GPU-accelerated compute-optimized instance family

Features of the gn7e instance family include:

  • Overview:

    • this instance family lets you select instance types with different numbers of GPUs and CPU resources to meet your various AI business needs.

    • Built on the third-generation X-Dragon architecture, gn7e instances deliver double the average network bandwidth for VPCs and cloud disks compared to the previous generation.

  • Use cases:

    • Small- and medium-scale AI training workloads.

    • High-performance computing (HPC) workloads accelerated using CUDA.

    • AI inference workloads that require high GPU compute performance or large GPU memory.

    • Deep learning, such as training AI algorithms for image classification, autonomous driving, and speech recognition.

    • GPU-intensive scientific computing, such as computational fluid dynamics, computational finance, molecular dynamics, and environmental analysis.

    Important

    When you run AI training workloads with high communication loads, such as those involving Transformer models, you must enable NVLink for GPU-to-GPU communication. Otherwise, large-scale data transfers over the PCIe link may cause unexpected failures and data corruption. If you are unsure about the communication link topology for your training workload, submit a ticket for support from Alibaba Cloud technical experts.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For more information, see IPv6 communication.

    • Network performance scales with the instance type. Larger instance types offer better network performance.

The gn7e instance family includes the instance types and specifications described in the following table.

Instance type

vCPUs

Memory (GiB)

GPU memory

Baseline bandwidth (Gbit/s)

Forwarding rate (pps)

Queues

ENIs

Private IPv4 addresses

IPv6 addresses

ecs.gn7e-c16g1.4xlarge

16

125

80 GB × 1

8

3,000,000

8

8

10

1

ecs.gn7e-c16g1.8xlarge

32

250

80 GB × 2

16

6,000,000

16

8

10

1

ecs.gn7e-c16g1.16xlarge

64

500

80 GB × 4

32

12,000,000

32

8

10

1

ecs.gn7e-c16g1.32xlarge

128

1000

80 GB × 8

64

24,000,000

32

16

15

1

gn7i, GPU-accelerated compute-optimized instance family

  • Overview: Powered by the third-generation SHENLONG architecture, gn7i instances deliver stable and predictable high performance. They use chip-level fast path acceleration to increase storage performance, network performance, and compute stability by an order of magnitude.

  • Use cases:

    • Equipped with high-performance CPUs, memory, and GPUs, these instances are ideal for concurrent AI inference tasks, such as image recognition, speech recognition, and behavior recognition.

    • These instances support RTX features and use high-frequency CPUs to deliver high-performance 3D graphics virtualization. They are suitable for graphics-intensive workloads, such as remote graphics design and cloud gaming.

  • Compute:

    • Equipped with NVIDIA A10 GPUs that feature:

      • The innovative NVIDIA Ampere architecture.

      • Support for common acceleration features such as RTX and TensorRT.

    • Processor: 2.9 GHz Intel ® Xeon ® Scalable (Ice Lake) processor with an all-core turbo frequency of 3.5 GHz.

    • This instance family provides up to 752 GiB of memory, a significant increase compared to the gn6i instance family.

  • Storage:

  • Network:

    • These instances support IPv4 and IPv6. For more information, see IPv6 communication.

    • Network performance scales with the instance type. Larger instance types offer better network performance.

The gn7i instance family includes the following instance types and specifications.

Instance type

vCPUs

Memory (GiB)

GPU

GPU memory

Network bandwidth (Gbit/s)

Packet rate (PPS)

NIC queues

ENIs

Private IPv4 addresses

IPv6 addresses

ecs.gn7i-c8g1.2xlarge

8

30

NVIDIA A10 * 1

24 GB * 1

16

1,600,000

8

4

15

15

ecs.gn7i-c16g1.4xlarge

16

60

NVIDIA A10 * 1

24 GB * 1

16

3,000,000

8

8

30

30

ecs.gn7i-c32g1.8xlarge

32

188

NVIDIA A10 * 1

24 GB * 1

16

6,000,000

12

8

30

30

ecs.gn7i-c32g1.16xlarge

64

376

NVIDIA A10 * 2

24 GB * 2

32

12,000,000

16

15

30

30

ecs.gn7i-c32g1.32xlarge

128

752

NVIDIA A10 * 4

24 GB * 4

64

24,000,000

32

15

30

30

ecs.gn7i-c48g1.12xlarge

48

310

NVIDIA A10 * 1

24 GB * 1

16

9,000,000

16

8

30

30

ecs.gn7i-c56g1.14xlarge

56

346

NVIDIA A10 * 1

24 GB * 1

16

10,000,000

16

8

30

30

ecs.gn7i-2x.8xlarge

32

128

NVIDIA A10 * 2

24 GB * 2

16

6,000,000

16

8

30

30

ecs.gn7i-4x.8xlarge

32

128

NVIDIA A10 * 4

24 GB * 4

32

6,000,000

16

8

30

30

ecs.gn7i-4x.16xlarge

64

256

NVIDIA A10 * 4

24 GB * 4

64

12,000,000

32

8

30

30

ecs.gn7i-8x.32xlarge

128

512

NVIDIA A10 * 8

24 GB * 8

64

24,000,000

32

16

30

30

ecs.gn7i-8x.16xlarge

64

256

NVIDIA A10 * 8

24 GB * 8

32

12,000,000

32

8

30

30

Important

You can change instances of the types ecs.gn7i-2x.8xlarge, ecs.gn7i-4x.8xlarge, ecs.gn7i-4x.16xlarge, ecs.gn7i-8x.32xlarge, and ecs.gn7i-8x.16xlarge to ecs.gn7i-c8g1.2xlarge or ecs.gn7i-c16g1.4xlarge. However, you cannot change them to other instance types such as ecs.gn7i-c32g1.8xlarge.

gn7s, GPU-accelerated compute-optimized instance family

To use the gn7s instance family, submit a ticket.

  • Introduction:

    • This instance family is powered by the latest Intel Ice Lake processors and NVIDIA A30 GPUs based on the NVIDIA Ampere architecture. This family offers various instance types with different GPU and CPU configurations to meet your specific AI needs.

    • Built on Alibaba Cloud's third-generation SHENLONG architecture, gn7s instances deliver twice the average network bandwidth for VPCs and cloud disks as the previous generation.

  • Use cases: Featuring high-performance CPUs, memory, and GPUs, these instances are ideal for concurrent AI inference workloads, such as image recognition, speech recognition, and behavior identification.

  • Compute:

    • Features NVIDIA A30 GPUs, which include:

      • The innovative NVIDIA Ampere architecture.

      • Support for the Multi-Instance GPU (MIG) feature and acceleration based on second-generation Tensor Cores for a wide range of workloads.

    • Processor: 2.9 GHz Intel ® Xeon ® Scalable (Ice Lake) processor with an all-core turbo frequency of 3.5 GHz.

    • Offers significantly more memory than the previous-generation instance family.

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For more information, see IPv6 communication.

    • Network performance scales with the instance type.

The gn7s instance family includes the following instance types and specifications:

Instance type

vCPUs

Memory (GiB)

GPUs

GPU memory

Network bandwidth (Gbit/s)

Packet rate (pps)

Private IPv4s per ENI

IPv6s per ENI

Multi-queue

ENIs

ecs.gn7s-c8g1.2xlarge

8

60

NVIDIA A30 * 1

24GB * 1

16

1,600,000

5

1

8

4

ecs.gn7s-c16g1.4xlarge

16

120

NVIDIA A30 * 1

24GB * 1

16

3,000,000

5

1

8

8

ecs.gn7s-c32g1.8xlarge

32

250

NVIDIA A30 * 1

24GB * 1

16

6,000,000

5

1

12

8

ecs.gn7s-c32g1.16xlarge

64

500

NVIDIA A30 * 2

24GB * 2

32

12,000,000

5

1

16

15

ecs.gn7s-c32g1.32xlarge

128

1000

NVIDIA A30 * 4

24GB * 4

64

24,000,000

10

1

32

15

ecs.gn7s-c48g1.12xlarge

48

380

NVIDIA A30 * 1

24GB * 1

16

9,000,000

8

1

16

8

ecs.gn7s-c56g1.14xlarge

56

440

NVIDIA A30 * 1

24GB * 1

16

10,000,000

8

1

16

8

gn7, GPU-accelerated compute-optimized instance family

  • Scenarios:

    • Deep learning, such as training AI algorithms used in image classification, autonomous driving, and speech recognition.

    • GPU-intensive scientific computing, such as computational fluid dynamics, computational finance, molecular dynamics, and environmental analysis.

The following table describes the instance types and specifications of the gn7 instance family.

Instance type

vCPUs

Memory (GiB)

GPU memory

Network bandwidth (Gbit/s)

Packet rate (pps)

NIC queues

ENIs

Private IPv4 addresses

IPv6 addresses

ecs.gn7-c12g1.3xlarge

12

94

40 GB × 1

4

2,500,000

4

8

10

1

ecs.gn7-c13g1.13xlarge

52

378

40 GB × 4

16

9,000,000

16

8

30

30

ecs.gn7-c13g1.26xlarge

104

756

40 GB × 8

30

18,000,000

16

15

10

1

gn6i, GPU-accelerated compute-optimized instance family

  • Use cases:

    • AI (deep learning and machine learning) inference for applications such as computer vision, speech recognition, speech synthesis, natural language processing (NLP), machine translation, and recommendation systems.

    • Real-time rendering for cloud gaming.

    • Real-time, cloud-based rendering for augmented reality (AR) and virtual reality (VR).

    • Graphics-heavy computing or graphics workstations.

    • GPU-accelerated databases.

    • High-performance computing (HPC).

  • Compute:

    • Equipped with NVIDIA T4 GPU accelerators, featuring:

      • The innovative NVIDIA Turing architecture.

      • 16 GB of memory per GPU with a memory bandwidth of 320 GB/s.

      • 2,560 CUDA cores per GPU.

      • Up to 320 Turing Tensor Cores per GPU.

      • Mixed-precision Tensor Cores that support 65 TFLOPS of FP16, 130 TOPS of INT8, and 260 TOPS of INT4.

    • vCPU-to-memory ratio of approximately 1:4.

    • Processor: 2.5 GHz Intel® Xeon® Platinum 8163 (Skylake).

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For details, see IPv6 communication.

    • Network performance scales with the instance type.

The gn6i instance family includes the following instance types.

Instance type

vCPUs

Memory (GiB)

GPUs

GPU memory

Network bandwidth (Gbit/s)

Packet rate (pps)

Disk IOPS

Multi-queue

ENIs

Private IPv4 addresses

IPv6 addresses

ecs.gn6i-c4g1.xlarge

4

15

NVIDIA T4 × 1

16 GB × 1

4

2,500,000

N/A

2

2

10

1

ecs.gn6i-c8g1.2xlarge

8

31

NVIDIA T4 × 1

16 GB × 1

5

2,500,000

N/A

2

2

10

1

ecs.gn6i-c16g1.4xlarge

16

62

NVIDIA T4 × 1

16 GB × 1

6

2,500,000

N/A

4

3

10

1

ecs.gn6i-c24g1.6xlarge

24

93

NVIDIA T4 × 1

16 GB × 1

7.5

2,500,000

N/A

6

4

10

1

ecs.gn6i-c40g1.10xlarge

40

155

NVIDIA T4 × 1

16 GB × 1

10

2,500,000

N/A

16

10

10

1

ecs.gn6i-c24g1.12xlarge

48

186

NVIDIA T4 × 2

16 GB × 2

15

4,500,000

N/A

12

6

10

1

ecs.gn6i-c24g1.24xlarge

96

372

NVIDIA T4 × 4

16 GB × 4

30

4,500,000

250,000

24

8

10

1

gn6e, GPU-accelerated compute-optimized instance family

  • Use cases:

    • Deep learning applications, such as training and inference for AI algorithms for image classification, autonomous driving, and speech recognition.

    • Scientific computing, such as computational fluid dynamics, computational finance, molecular dynamics, and environmental analysis.

  • Compute:

    • Features NVIDIA V100 (32 GB NVLink) GPU cards.

    • GPU accelerator: V100 (SXM2 package).

      • Innovative NVIDIA Volta architecture.

      • 32 GB of HBM2 memory per GPU with a GPU memory bandwidth of 900 GB/s.

      • 5,120 CUDA Cores per GPU.

      • 640 Tensor Cores per GPU.

      • Each GPU supports six bidirectional NVLink connections, each providing 25 Gbit/s of bandwidth in each direction for a total of 300 Gbit/s.

    • Features a vCPU-to-memory ratio of approximately 1:8.

    • Processor: 2.5 GHz Intel ® Xeon ® Platinum 8163 (Skylake).

  • Storage:

  • Network:

    • Supports both IPv4 and IPv6. For more information about IPv6 communication, see IPv6 communication.

    • Network performance scales with the instance type.

gn6e includes the instance types and specifications listed in the table below.

Instance type

vCPU

Memory (GiB)

GPU

GPU memory

Baseline bandwidth (Gbit/s)

Packet rate (PPS)

NIC queues

ENI

Private IPv4 addresses

IPv6 addresses

ecs.gn6e-c12g1.3xlarge

12

92

1 × NVIDIA V100

1 × 32 GB

5

800,000

8

6

10

1

ecs.gn6e-c12g1.6xlarge

24

184

2 × NVIDIA V100

2 × 32 GB

8

1,200,000

8

8

20

1

ecs.gn6e-c12g1.12xlarge

48

368

4 × NVIDIA V100

4 × 32 GB

16

2,400,000

8

8

20

1

ecs.gn6e-c12g1.24xlarge

96

736

8 × NVIDIA V100

8 × 32 GB

32

4,500,000

16

8

20

1

gn6v, GPU-accelerated compute-optimized instance family

  • Use cases:

    • Deep learning applications, such as training and inference for AI algorithms in image classification, autonomous driving, and speech recognition.

    • Scientific computing, such as computational fluid dynamics, computational finance, molecular dynamics, and environmental analysis.

  • Compute:

    • Equipped with NVIDIA V100 GPUs.

    • GPU accelerator: V100 (SXM2 package).

      • Innovative NVIDIA Volta architecture.

      • 16 GB of HBM2 GPU memory per GPU with 900 GB/s of memory bandwidth.

      • 5,120 CUDA Cores per GPU.

      • 640 Tensor Cores per GPU.

      • Up to six NVLink bidirectional connections per GPU. Each connection provides a bandwidth of 25 Gbit/s in each direction, for a total bandwidth of 300 Gbit/s.

    • Features a vCPU-to-memory ratio of approximately 1:4.

    • Processor: 2.5 GHz Intel® Xeon® Platinum 8163 (Skylake).

  • Storage:

  • Network:

    • Supports IPv4 and IPv6. For more information, see IPv6 communication.

    • Network performance scales with the instance type.

The gn6v instance family includes the instance types and specifications listed below.

Instance type

vCPU

Memory (GiB)

GPU

GPU memory

Network bandwidth (Gbit/s)

Packet rate (pps)

Disk baseline IOPS

Multi-queue

ENI

Private IPv4 addresses

IPv6 addresses

ecs.gn6v-c8g1.2xlarge

8

32

1 × NVIDIA V100

1 × 16 GB

2.5

800,000

N/A

4

4

10

1

ecs.gn6v-c8g1.4xlarge

16

64

2 × NVIDIA V100

2 × 16 GB

5

1,000,000

N/A

4

8

20

1

ecs.gn6v-c8g1.8xlarge

32

128

4 × NVIDIA V100

4 × 16 GB

10

2,000,000

N/A

8

8

20

1

ecs.gn6v-c8g1.16xlarge

64

256

8 × NVIDIA V100

8 × 16 GB

20

2,500,000

N/A

16

8

20

1

ecs.gn6v-c10g1.20xlarge

82

336

8 × NVIDIA V100

8 × 16 GB

35

4,500,000

250,000

16

8

20

1