All Products
Search
Document Center

Tair:Persistent memory-optimized instances

Last Updated:Jun 05, 2023

This topic describes the specifications of Tair persistent memory-optimized instances. These specifications include the memory capacity, maximum number of connections to each instance, and maximum bandwidth.

Instance specifications

The following tables describe the specifications of a single shard. You can calculate the specifications of a cluster or read/write splitting instance by using the following formula: Shard specifications × Number of shards.

Note
  • Enhanced SSDs (ESSDs) described in the following table are provided for dedicated storage of system files, such as logs, backups of temporary files, and append-only files (AOFs). For more information about ESSDs, see ESSDs.
  • For information about the cluster architecture in direct connection mode and that in proxy mode, see Cluster architecture.
  • Cluster instances that use cloud disks can have a custom number of shards. For more information, see Adjust the number of data shards for a Tair instance with cloud disks.

Shard specifications for cloud disk-based standard instances and cluster instances in direct connection mode

The specifications listed in the following table are applicable to cloud disk-based standard and cluster instances in direct connection mode.

Table 1. Shard specifications for cloud disk-based standard instances and cluster instances in direct connection mode

Instance type

InstanceClass value (used in API operations)

Specifications

Maximum connections

Maximum bandwidth (MB/s)

QPS reference value

Number of CPU cores

Persistent memory (GB)

ESSD capacity (GB)

4 GB

tair.scm.standard.1m.4d

3

4

12

10,000

96

100,000

8 GB

tair.scm.standard.2m.8d

3

8

24

10,000

96

100,000

16 GB

tair.scm.standard.4m.16d

3

16

48

10,000

96

100,000

32 GB

tair.scm.standard.8m.32d

3

32

96

10,000

96

100,000

64 GB

tair.scm.standard.16m.64d

3

64

192

10,000

96

100,000

Note

The 4 GB instance type is unavailable for cluster instances in direct connection mode.

Shard specifications of cluster instances in proxy mode and read/write splitting instances

The specifications listed in the following table are applicable to cloud disk-based cluster instances in proxy mode and cloud disk-based read/write splitting instances.

Table 2. Shard specifications of cluster instances in proxy mode and read/write splitting instances

Instance type

InstanceClass value (used in API operations)

Specifications

Maximum connections

Maximum bandwidth (MB/s)

QPS reference value

Number of CPU cores

Persistent memory (GB)

ESSD capacity (GB)

8 GB

tair.scm.with.proxy.standard.2m.8d

3

8

24

10,000

96

100,000

16 GB

tair.scm.with.proxy.standard.4m.16d

3

16

48

10,000

96

100,000

32 GB

tair.scm.with.proxy.standard.8m.32d

3

32

96

10,000

96

100,000

64 GB

tair.scm.with.proxy.standard.16m.64d

3

64

192

10,000

96

100,000

Performance description

Architecture

Description

Standard architecture

The overall performance of standard instances is as described in Table 1.

Cluster architecture in direct connection mode

Cluster architecture in proxy mode

Read/write splitting architecture

Instance performance = Shard specifications × Number of shards.

Note When a cluster instance that runs in proxy mode is accessed, the maximum bandwidth is 2,048 Mbit/s, the maximum total QPS is 10,000,000, and the maximum number of connections is 500,000.

For example, assume that an instance has four shards of the tair.scm.with.proxy.standard.2m.8d type with the following specifications:

  • CPU: 6 cores

  • Bandwidth: 96 MB/s

  • Maximum connections: 30,000

In this case, the instance provides the following performance:

  • CPU: 24 cores

  • Bandwidth: 384 MB/s

  • Maximum connections: 120,000

Notes about CPU cores

To ensure service stability, the system reserves a CPU core to process background tasks. In a cluster instance or a read/write splitting instance, the system reserves a CPU core for each data shard or each read replica to process background tasks.

Bandwidth description

  • The bandwidth described in the preceding table is that of a single shard. The overall bandwidth of a cluster or read/write splitting instance is calculated by using the following formula: Shard bandwidth × Number of shards.
  • The upper limit for cluster instances in proxy mode is 2,048 Mbit/s. If the limit is reached, the bandwidth of an instance cannot be increased by adding shards. Cluster instances in direct connection mode have no limits on the overall bandwidth.
  • The bandwidth value applies to the upstream and downstream bandwidths. For example, if the overall bandwidth of an instance is 24 Mbit/s, the upstream and downstream bandwidths of the instance are both 24 Mbit/s.
  • The bandwidth in the preceding tables represents the internal bandwidth of Tair instances. The public bandwidth is determined by the internal bandwidth and is limited by the bandwidth of the connection between a Tair instance and the client. We recommend that you connect to the instance over an internal network to maximize performance.

Calculation rules for connections

ItemDescription
Maximum number of concurrent connections
  • Proxy mode: The maximum number of connections to an instance is 500,000. If the upper limit of 500,000 connections is reached, subsequent connections cannot be established even if you add more shards or nodes.
  • Direct connection mode: The maximum number of connections for a single shard is 10,000. The maximum number of connections for an instance can be calculated by using the following formula: Number of shards × 10,000.
Number of new connections per secondThe upper limit of the number of new connections per second refers to the number of connections that are allowed to be established per second. For example, assume that the maximum number of new connections to an instance per second is 10,000, the maximum number of connections to the instance is 50,000, and the actual number of connections at the Nth second after the instance starts to run is 12,000. In this case, at the (N+1)th second, the maximum number of connections is 22,000. This value is calculated by using the following formula: 12,000 + 10,000.
Note If the instance is connected in direct connection mode, connection performance degrades because proxy nodes are bypassed. The maximum number of new connections per second for a single data shard is 2,000. For example, assume that an instance has four data shards, the maximum number of new connections per second is 8,000.