what are local disks and local disk categories and how are local disks used - Elastic Compute Service

Local disks reside on the physical machines that host the associated Elastic Compute Service (ECS) instances of the disks, and provide ECS instances access to local storage. Local disks are physically attached to physical machines and deliver cost-effectiveness, high random IOPS, high throughput, and low latency. Local disks are suitable for scenarios that require high-performance storage, such as cache systems and databases that involve a large number of random accesses. This topic describes the categories and performance specifications of local disks.

Limits

All of the local disks for an instance reside on a single physical machine. This increases single point of failure (SPOF) risks. The durability of data stored on a local disk is determined by the reliability of the associated physical machine.
Warning
For example, data stored on local disks may be lost when a hardware failure occurs on their associated physical machine. We recommend that you store only temporary data on local disks.
- To ensure data availability, we recommend that you implement data redundancy at the application layer. You can use deployment sets to distribute ECS instances across multiple physical machines for high availability and disaster recovery purposes. For more information, see Create a deployment set.
- If your applications do not use an architecture that prioritizes data reliability, we recommend that you use cloud disks or a backup service with ECS instances to improve data reliability. For more information, see Disks or What is Cloud Backup?
After you purchase an ECS instance that has local disks attached, log on to the instance to partition and format the local disks. For more information, see Initialize a data disk whose size does not exceed 2 TiB on a Linux instance or Initialize a data disk up to 2 TiB in size on a Windows instance.
Local disks do not support the following operations:
- Create a separate local disk.
- Use a snapshot to create a local disk.
- Attach a local disk.
- Detach and release a local disk.
- Resize a local disk.
- Re-initialize a local disk.
- Create a snapshot for a local disk.
- Use a snapshot to roll back a local disk.

Local disk categories

Note

This topic provides information about local disks that are purchased together with ECS instances. For information about the performance of instance families that are equipped with local SSDs and big data instance families, see Overview of instance families.

Local disks are suitable for scenarios that require high storage I/O performance, mass storage, and high cost efficiency. Alibaba Cloud provides two categories of local disks. The following table describes the categories.

Category	Supported instance family	Scenario
Local non-volatile memory express (NVMe) SSD	The following instance families use local NVMe SSDs: Instance families equipped with local SSDs: i4, i4g, i4r, i3, i3g, i2, i2g, i2ne, i2gne, and i1 GPU-accelerated compute-optimized instance family: gn5	Instance families equipped with local NVMe SSDs are suitable for the following scenarios: I/O-intensive applications that require high I/O performance and low latency, such as online gaming, e-commerce, live streaming, and media Applications that require high storage I/O performance and a high-availability architecture at the application layer, such as NoSQL databases (including Cassandra, MongoDB, and HBase), Massively Parallel Processing (MPP) data warehouses, and distributed file systems
Local SATA HDD	The d3s, d2c, d2s, d1ne, and d1 big data instance families use local SATA HDDs.	Local SATA HDDs are the preferred storage media for industries such as Internet and finance that have high requirements for big data computing, storage, and analytics. These disks are suited for mass storage and offline computing scenarios and can meet the high requirements of distributed computing services such as Hadoop in terms of storage performance, storage capacity, and internal network bandwidth.

Performance of local disks

Local NVMe SSDs

The following table describes the performance metrics of local NVMe SSDs that the d3c compute-intensive big data instance family uses.

Metric	Single-disk performance	ecs.d3c.3xlarge	ecs.d3c.7xlarge	ecs.d3c.14xlarge
Maximum read IOPS	100,000	100,000	200,000	400,000
Maximum read throughput	4 GB/s	4 GB/s	8 GB/s	16 GB/s
Maximum write throughput	2 GB/s	2 GB/s	4 GB/s	8 GB/s

The following table describes the performance metrics of the local NVMe SSDs that the i4 instance family uses.

Metric	ecs.i4.large	ecs.i4.xlarge	ecs.i4.2xlarge	ecs.i4.4xlarge	ecs.i4.8xlarge	ecs.i4.16xlarge	ecs.i4.32xlarge
Maximum read IOPS	112,500	225,000	450,000	900,000	1,800,000	3,600,000	7,200,000
Maximum read throughput	0.75 GB/s	1.5 GB/s	3 GB/s	6 GB/s	12 GB/s	24 GB/s	48 GB/s
Maximum write throughput	0.375 GB/s	0.75 GB/s	1.5 GB/s	3 GB/s	6 GB/s	12 GB/s	24 GB/s

Note

The performance data in the preceding table represents the highest performance levels of local storage for the instance families. The instance families support only Linux images. We recommend that you use the most recent Linux image versions, such as Alibaba Cloud Linux 3, to obtain optimal performance.

The following table describes the performance metrics of the local NVMe SSDs that the i4g and i4r instance families use.

Metric	ecs.i4g.4xlarge and ecs.i4r.4xlarge	ecs.i4g.8xlarge and ecs.i4r.8xlarge	ecs.i4g.16xlarge and ecs.i4r.16xlarge	ecs.i4g.32xlarge and ecs.i4r.32xlarge
Maximum read IOPS	250,000	500,000	1,000,000	2,000,000
Maximum read throughput	1.5 GB/s	3 GB/s	6 GB/s	12 GB/s
Maximum write throughput	1 GB/s	2 GB/s	4 GB/s	8 GB/s

Note

The following table describes the performance metrics of the local NVMe SSDs that the i3 instance family uses.

Metric	ecs.i3.xlarge	ecs.i3.2xlarge	ecs.i3.4xlarge	ecs.i3.8xlarge	ecs.i3.13xlarge	ecs.i3.26xlarge
Maximum read IOPS	250,000	500,000	1,000,000	2,000,000	3,000,000	6,000,000
Maximum read throughput	1.5 GB/s	3 GB/s	6 GB/s	12 GB/s	18 GB/s	36 GB/s
Maximum write throughput	1 GB/s	2 GB/s	4 GB/s	8 GB/s	12 GB/s	24 GB/s

Note

The following table describes the performance metrics of the local NVMe SSDs that the i3g instance family uses.

Metric	ecs.i3g.2xlarge	ecs.i3g.4xlarge	ecs.i3g.8xlarge	ecs.i3g.13xlarge	ecs.i3g.26xlarge
Maximum read IOPS	125,000	250,000	500,000	750,000	1,500,000
Maximum read throughput	0.75 GB/s	1.5 GB/s	3 GB/s	4.5 GB/s	9 GB/s
Maximum write throughput	0.5 GB/s	1 GB/s	2 GB/s	3 GB/s	6 GB/s

Note

The following table describes the performance metrics of the local NVMe SSDs that the i2 and i2g instance families use.

Metric	Single-disk performance		Overall instance performance^①
Metric	ecs.i2.xlarge and ecs.i2g.2xlarge	Other i2 and i2g instance types	Overall instance performance^①
Maximum capacity	894 GiB	1,788 GiB	8 × 1,788 GiB
Maximum read IOPS	150,000	300,000	1,500,000
Maximum read throughput	1 GB/s	2 GB/s	16 GB/s
Maximum write throughput	0.5 GB/s	1 GB/s	8 GB/s
Access latency	Several microseconds

^① The overall instance performance data in the preceding table applies only to the ecs.i2.16xlarge instance type and represents the highest performance levels of local storage for the i2 instance family.

The following table describes the performance metrics of the local NVMe SSDs that the i2ne and i2gne instance families use.

Metric	ecs.i2ne.xlarge and ecs.i2gne.2xlarge	ecs.i2ne.2xlarge and ecs.i2gne.4xlarge	ecs.i2ne.4xlarge and ecs.i2gne.8xlarge	ecs.i2ne.8xlarge and ecs.i2gne.16xlarge	ecs.i2ne.16xlarge
Maximum capacity	894 GiB	1,788 GiB	2 × 1,788 GiB	4 × 1,788 GiB	8 × 1,788 GiB
Maximum read IOPS	250,000	500,000	1,000,000	2,000,000	4,000,000
Maximum read throughput	1.5 GB/s	3 GB/s	6 GB/s	12 GB/s	24 GB/s
Maximum write throughput	1 GB/s	2 GB/s	4 GB/s	8 GB/s	16 GB/s
Access latency	Several microseconds

The following table describes the performance metrics of the local NVMe SSDs that the i1 instance family uses.

Metric	Single-disk performance	Overall instance performance^②
Maximum capacity	1,456 GiB	2,912 GiB
Maximum IOPS	240,000	480,000
Write IOPS ^①	min{165 × Capacity, 240,000}	2 × min{165 × Capacity, 240,000}
Read IOPS ^①	min{165 × Capacity, 240,000}	2 × min{165 × Capacity, 240,000}
Maximum read throughput	2 GB/s	4 GB/s
Read throughput ^①	min{1.4 × Capacity, 2,000} MB/s	2 × min{1.4 × Capacity, 2,000} MB/s
Maximum write throughput	1.2 GB/s	2.4 GB/s
Write throughput ^①	min{0.85 × Capacity, 1,200} MB/s	2 × min{0.85 × Capacity, 1,200} MB/s
Access latency	Several microseconds

^① Items in the formulas used to calculate the performance specifications of a single local NVMe SSD.

In the formula used to calculate the write IOPS, each GiB of capacity produces 165 write IOPS for up to 240,000 IOPS per disk.
In the formula used to calculate the write throughput, each GiB of capacity produces a write throughput of 0.85 MB/s for a maximum throughput of 1,200 MB/s per disk.

^② The overall instance performance data in the preceding table applies only to the ecs.i1.14xlarge instance type and represents the highest performance levels of local storage for the i1 instance family.

Local SATA HDDs

The following table describes the performance metrics of local SATA HDDs.

Metric	d1 and d1ne		d2c		d2s		d3s
Metric	Single-disk performance	Overall instance performance	Single-disk performance	Overall instance performance	Single-disk performance	Overall instance performance	Single-disk performance	Overall instance performance
Maximum capacity	5,500 GiB	154,000 GiB	3,700 GiB	44,400 GiB	7,300 GiB	219,000 GiB	11,100 GiB	355,200 GiB
Maximum throughput	190 MB/s	5,320 MB/s	190 MB/s	2,280 MB/s	190 MB/s	5,700 MB/s	260 MB/s	8,320 MB/s
Access latency	Several milliseconds

Note

The overall instance performance data in the preceding table applies only to the ecs.d1.14xlarge, ecs.d1ne.14xlarge, ecs.d2c.24xlarge, ecs.d2s.20xlarge, and ecs.d3s.16xlarge instance types and represents the highest performance levels of local storage for the corresponding instance families.

Billing

Fees for local disks are included in the fees for the instances to which the disks are attached. For more information, see Subscription and Pay-as-you-go.

Disk initialization sequence

When you use an image to create an instance that has local disks attached, disks on the created instance are initialized based on the following rules:

Rule 1: If the image does not contain data disk snapshots, the local disks are initialized prior to the cloud disks that are created together with the instance.
Rule 2: If the image contains data disk snapshots, the data disks created from the snapshots are initialized in the order of data disk device names that are recorded in the image. The other disks on the instance are initialized based on Rule 1.

The following section provides an example on how disks are initialized based on Rule 2. In this example, an instance created from a Linux image that contains two data disk snapshots is used.

If the data disk device names recorded in the image are /dev/xvdb and /dev/xvdc, Alibaba Cloud first allocates /dev/xvdb and /dev/xvdc as device names to the data disks created from the image. The system disk is initialized first. Then, the data disks are initialized in the following sequence: data disk 1 created from the image, data disk 2 created from the image, local disk 1, local disk 2, cloud disk 1, cloud disk 2, and so on. The following figure shows the sequence in which the disks are initialized.
If the data disk device names recorded in the image are /dev/xvdc and /dev/xvdd, Alibaba Cloud first allocates /dev/xvdc and /dev/xvdd as device names to the data disks created from the image. Then, Alibaba Cloud allocates other available device names in alphabetic order to the local disks first and then to other disks. The system disk is initialized first. Then, the data disks are initialized in the following sequence: local disk 1, data disk 1 created from the image, data disk 2 created from the image, local disk 2, cloud disk 1, cloud disk 2, and so on. The following figure shows the sequence in which the disks are initialized.

Lifecycle

A local disk shares the same lifecycle as the instance to which it is attached. For more information, see Instance lifecycle.

Impacts of instance operations on data stored on local disks

The following table describes the impacts of instance operations on data stored on local disks.

Instance operation	Data stored on local disks	Local disk
Restart the operating system, restart an instance in the ECS console, or forcefully restart an instance.	Retained	Retained
Shut down the operating system, stop an instance in the ECS console, or forcefully stop an instance.	Retained	Retained
Automatically recover an instance.	Erased	Released
Release an instance	Erased	Released
A subscription instance is stopped on expiration and has not been released, or a pay-as-you-go instance is stopped due to an overdue payment and has not been released.	Retained	Retained
A subscription instance is stopped on expiration and then released, or a pay-as-you-go instance is stopped due to an overdue payment and then released.	Erased	Released
Manually renew an expired subscription instance.	Retained	Retained
Reactivate a pay-as-you-go instance that is stopped due to an overdue payment.	Retained	Retained

References

You can test the bandwidth, IOPS, and latency of local disks to verify the benchmark performance data provided by Alibaba Cloud and the Quality of Service (QoS) of local disks. For more information, see the Commands used to test the performance of local disks section of the "Test the performance of block storage devices" topic.
For information about retired local SSDs, see Previous-generation disks - local SSDs.
For information about how to handle system events of instances that are equipped with local disks, see O&M scenarios and system events for instances equipped with local disks.