Supported node types
ElastiCache supports the following node types. Generally speaking, the current generation types provide more memory and computational power at lower cost when compared to their equivalent previous generation counterparts.
For more information on performance details for each node type, see Amazon EC2 Instance Types
For information on which node size to use, see Choosing your node size.
Current Generation
For more information on Previous Generation, please refer to Previous
Generation Nodes
Note
Instance types with burstable network performance use a network I/O credit mechanism to burst beyond their baseline bandwidth on a best-effort basis.
General
| Instance type | Minimum supported Redis OSS version | Enhanced I/O (Redis OSS 5.0.6+) | TLS Offloading (Redis OSS 6.2.5+) | Enhanced I/O Multiplexing (Redis OSS 7.0.4+) | Baseline bandwidth (Gbps) | Burst bandwidth (Gbps) |
|---|---|---|---|---|---|---|
| cache.m7g.large | 6.2 | N | N | N | 0.937 | 12.5 |
| cache.m7g.xlarge | 6.2 | Y | Y | Y | 1.876 | 12.5 |
| cache.m7g.2xlarge | 6.2 | Y | Y | Y | 3.75 | 15 |
| cache.m7g.4xlarge | 6.2 | Y | Y | Y | 7.5 | 15 |
| cache.m7g.8xlarge | 6.2 | Y | Y | Y | 15 | N/A |
| cache.m7g.12xlarge | 6.2 | Y | Y | Y | 22.5 | N/A |
| cache.m7g.16xlarge | 6.2 | Y | Y | Y | 30 | N/A |
| cache.m6g.large | 5.0.6 | N | N | N | 0.75 | 10.0 |
| cache.m6g.xlarge | 5.0.6 | Y | Y | Y | 1.25 | 10.0 |
| cache.m6g.2xlarge | 5.0.6 | Y | Y | Y | 2.5 | 10.0 |
| cache.m6g.4xlarge | 5.0.6 | Y | Y | Y | 5.0 | 10.0 |
| cache.m6g.8xlarge | 5.0.6 | Y | Y | Y | 12 | N/A |
| cache.m6g.12xlarge | 5.0.6 | Y | Y | Y | 20 | N/A |
| cache.m6g.16xlarge | 5.0.6 | Y | Y | Y | 25 | N/A |
| cache.m5.large | 3.2.4 | N | N | N | 0.75 | 10.0 |
| cache.m5.xlarge | 3.2.4 | Y | N | N | 1.25 | 10.0 |
| cache.m5.2xlarge | 3.2.4 | Y | Y | Y | 2.5 | 10.0 |
| cache.m5.4xlarge | 3.2.4 | Y | Y | Y | 5.0 | 10.0 |
| cache.m5.12xlarge | 3.2.4 | Y | Y | Y | 12 | N/A |
| cache.m5.24xlarge | 3.2.4 | Y | Y | Y | 25 | N/A |
| cache.m4.large | 3.2.4 | N | N | N | 0.45 | 1.2 |
| cache.m4.xlarge | 3.2.4 | Y | N | N | 0.75 | 2.8 |
| cache.m4.2xlarge | 3.2.4 | Y | Y | Y | 1.0 | 10.0 |
| cache.m4.4xlarge | 3.2.4 | Y | Y | Y | 2.0 | 10.0 |
| cache.m4.10xlarge | 3.2.4 | Y | Y | Y | 5.0 | 10.0 |
| cache.t4g.micro | 3.2.4 | N | N | N | 0.064 | 5.0 |
| cache.t4g.small | 5.0.6 | N | N | N | 0.128 | 5.0 |
| cache.t4g.medium | 5.0.6 | N | N | N | 0.256 | 5.0 |
| cache.t3.micro | 3.2.4 | N | N | N | 0.064 | 5.0 |
| cache.t3.small | 3.2.4 | N | N | N | 0.128 | 5.0 |
| cache.t3.medium | 3.2.4 | N | N | N | 0.256 | 5.0 |
| cache.t2.micro | 3.2.4 | N | N | N | 0.064 | 1.024 |
| cache.t2.small | 3.2.4 | N | N | N | 0.128 | 1.024 |
| cache.t2.medium | 3.2.4 | N | N | N | 0.256 | 1.024 |
Memory optimized
| Instance type | Minimum supported Redis OSS version | Enhanced I/O (Redis OSS 5.0.6+) | TLS Offloading (Redis OSS 6.2.5+) | Enhanced I/O Multiplexing (Redis OSS 7.0.4+) | Baseline bandwidth (Gbps) | Burst bandwidth (Gbps) |
|---|---|---|---|---|---|---|
| cache.r7g.large | 6.2 | N | N | N | 0.937 | 12.5 |
| cache.r7g.xlarge | 6.2 | Y | Y | Y | 1.876 | 12.5 |
| cache.r7g.2xlarge | 6.2 | Y | Y | Y | 3.75 | 15 |
| cache.r7g.4xlarge | 6.2 | Y | Y | Y | 7.5 | 15 |
| cache.r7g.8xlarge | 6.2 | Y | Y | Y | 15 | N/A |
| cache.r7g.12xlarge | 6.2 | Y | Y | Y | 22.5 | N/A |
| cache.r7g.16xlarge | 6.2 | Y | Y | Y | 30 | N/A |
| cache.r6g.large | 5.0.6 | N | N | N | 0.75 | 10.0 |
| cache.r6g.xlarge | 5.0.6 | Y | Y | Y | 1.25 | 10.0 |
| cache.r6g.2xlarge | 5.0.6 | Y | Y | Y | 2.5 | 10.0 |
| cache.r6g.4xlarge | 5.0.6 | Y | Y | Y | 5.0 | 10.0 |
| cache.r6g.8xlarge | 5.0.6 | Y | Y | Y | 12 | N/A |
| cache.r6g.12xlarge | 5.0.6 | Y | Y | Y | 20 | N/A |
| cache.r6g.16xlarge | 5.0.6 | Y | Y | Y | 25 | N/A |
| cache.r5.large | 3.2.4 | N | N | N | 0.75 | 10.0 |
| cache.r5.xlarge | 3.2.4 | Y | N | N | 1.25 | 10.0 |
| cache.r5.2xlarge | 3.2.4 | Y | Y | Y | 2.5 | 10.0 |
| cache.r5.4xlarge | 3.2.4 | Y | Y | Y | 5.0 | 10.0 |
| cache.r5.12xlarge | 3.2.4 | Y | Y | Y | 12 | N/A |
| cache.r5.24xlarge | 3.2.4 | Y | Y | Y | 25 | N/A |
| cache.r4.large | 3.2.4 | N | N | N | 0.75 | 10.0 |
| cache.r4.xlarge | 3.2.4 | Y | N | N | 1.25 | 10.0 |
| cache.r4.2xlarge | 3.2.4 | Y | Y | Y | 2.5 | 10.0 |
| cache.r4.4xlarge | 3.2.4 | Y | Y | Y | 5.0 | 10.0 |
| cache.r4.8xlarge | 3.2.4 | Y | Y | Y | 12 | N/A |
| cache.r4.16xlarge | 3.2.4 | Y | Y | Y | 25 | N/A |
Memory optimized with data tiering
| Instance type | Minimum supported Redis OSS version | Enhanced I/O (Redis OSS 5.0.6+) | TLS Offloading (Redis OSS 6.2.5+) | Enhanced I/O Multiplexing (Redis OSS 7.0.4+) | Baseline bandwidth (Gbps) | Burst bandwidth (Gbps) |
|---|---|---|---|---|---|---|
| cache.r6gd.xlarge | 6.2.0 | Y | N | N | 1.25 | 10 |
| cache.r6gd.2xlarge | 6.2.0 | Y | Y | Y | 2.5 | 10 |
| cache.r6gd.4xlarge | 6.2.0 | Y | Y | Y | 5.0 | 10 |
| cache.r6gd.8xlarge | 6.2.0 | Y | Y | Y | 12 | N/A |
| cache.r6gd.12xlarge | 6.2.0 | Y | Y | Y | 20 | N/A |
| cache.r6gd.16xlarge | 6.2.0 | Y | Y | Y | 25 | N/A |
Network optimized
| Instance type | Minimum supported Redis OSS version | Enhanced I/O (Redis OSS 5.0.6+) | TLS Offloading (Redis OSS 6.2.5+) | Enhanced I/O Multiplexing (Redis OSS 7.0.4+) | Baseline bandwidth (Gbps) | Burst bandwidth (Gbps) |
|---|---|---|---|---|---|---|
| cache.c7gn.large | 6.2 | N | N | N | 6.25 | 30 |
| cache.c7gn.xlarge | 6.2 | Y | Y | Y | 12.5 | 40 |
| cache.c7gn.2xlarge | 6.2 | Y | Y | Y | 25 | 50 |
| cache.c7gn.4xlarge | 6.2 | Y | Y | Y | 50 | N/A |
| cache.c7gn.8xlarge | 6.2 | Y | Y | Y | 100 | N/A |
| cache.c7gn.12xlarge | 6.2 | Y | Y | Y | 150 | N/A |
| cache.c7gn.16xlarge | 6.2 | Y | Y | Y | 200 | N/A |
Previous Generation Nodes support
Ending support for T1, M1, M2, M3, and R3
ElastiCache is ending support for T1, M1, M2, M3, R3 cache node types. Amazon Web Services offers Previous Generation Nodes for users who have optimized their applications around these nodes and have yet to upgrade. Previous Generation on-demand nodes will reach End-of-Life on October 15, 2024 in the Beijing (BJS) region operated by Sinnet and Ningxia (ZHY) region operated by NWCD. We recommend that you migrate these nodes to the newer generation node types before October 15, 2024.
Upgrade paths
We encourage you to use the latest generation of nodes to get the best performance, but we will continue to support Previous Generation Nodes after new nodes launch. If you are currently using a Previous Generation Node and would like to see a recommended upgrade path, see the table below.
M1/M3 to M5, M6 or M7
M5 instances provide the latest Intel Xeon or Arm processors. With M5, you get a higher performing CPU with support for AVX-512 instructions, up to 384 GiB of RAM, and up to 25 Gbps of networking bandwidth, all at a lower price point then our previous generation instances. Amazon EC2 M6g/M7g instances are powered by Arm-based Amazon Web Services Graviton processors. Graviton instances offer a balance of compute, memory, and networking resources for a broad set of workloads.
| Previous Generation Node type | Recommended Node type |
|---|---|
| cache.m1.small | cache.t3.small |
| cache.m1.medium | cache.t3.medium |
| cache.m1.large | cache.m5.large |
| cache.m1.xlarge | cache.m5.xlarge |
| cache.m2.2xlarge | cache.m5.2xlarge |
| cache.m2.4xlarge | cache.m6g.4xlarge |
| cache.m3.medium | cache.t3.medium |
| cache.m3.large | cache.m5.large |
| cache.m3.xlarge | cache.m5.xlarge |
| cache.m3.2xlarge | cache.m5.2xlarge |
M2/R3 to R5, R6, or R7
The R5 family of instances provide a newer Intel Xeon processor, up to 636.61 GiB of RAM, are EBS optimized, and support Enhanced Networking with up to 25 Gbps offering lower latency and low jitter. Amazon EC2 R6g/R7g instances are powered by Amazon Web Services Graviton processors (currently Graviton3 instances are not available in all Amazon Web Services Regions). Graviton instances are ideal for running memory-intensive workloads.
| Previous Generation Node type | Recommended Node type |
|---|---|
| cache.r3.large | cache.r5.large |
| cache.r3.xlarge | cache.r5.xlarge |
| cache.r3.2xlarge | cache.r5.2xlarge |
| cache.r3.4xlarge | cache.r5.4xlarge |
| cache.r3.8xlarge | cache.r6g.8xlarge |
T1/M1 to T3 or T4g
T3 and T4g nodes are Burstable Performance Nodes that provide a baseline level of CPU performance with the ability to burst above the baseline. They are recommended for workloads that don't use the full CPU often or consistently, but occasionally need to burst to higher CPU performance, such as test and development environments, as well as small workloads. Both T3 and T4g nodes are less expensive than m1.small, m1.medium, and t1.micro nodes in most Regions.
| Previous Generation Node type | Recommended Node type |
|---|---|
| cache.t1.micro | cache.t3.micro |
For more information on Previous Generation, please refer to Previous
Generation Nodes
Supported node types by Amazon Region
Supported node types may vary between Amazon Regions. For more details, see Amazon ElastiCache
pricing
Burstable Performance Instances
You can launch general-purpose burstable T4g, T3-Standard and T2-Standard cache nodes in Amazon ElastiCache. These nodes provide a baseline level of CPU performance with the ability to burst CPU usage at any time until the accrued credits are exhausted. A CPU credit provides the performance of a full CPU core for one minute.
Amazon ElastiCache's T4g, T3 and T2 nodes are configured as standard and suited for workloads with an average CPU utilization that is consistently below the baseline performance of the instance. To burst above the baseline, the node spends credits that it has accrued in its CPU credit balance. If the node is running low on accrued credits, performance is gradually lowered to the baseline performance level. This gradual lowering ensures the node doesn't experience a sharp performance drop-off when its accrued CPU credit balance is depleted. For more information, see CPU Credits and Baseline Performance for Burstable Performance Instances in the Amazon EC2 User Guide.
The following table lists the burstable performance node types, the rate at which CPU credits are earned per hour. It also shows the maximum number of earned CPU credits that a node can accrue and the number of vCPUs per node. In addition, it gives the baseline performance level as a percentage of a full core performance (using a single vCPU).
| Node type | CPU credits earned per hour | Maximum earned credits that can be accrued* | vCPUs | Baseline performance per vCPU | Memory (GiB) | Network performance |
|---|---|---|---|---|---|---|
| t4g.micro | 12 |
288 | 2 | 10% | 0.5 | Up to 5 Gigabit |
| t4g.small | 24 |
576 | 2 | 20% | 1.37 | Up to 5 Gigabit |
| t4g.medium | 24 |
576 | 2 | 20% | 3.09 | Up to 5 Gigabit |
| t3.micro | 12 |
288 | 2 | 10% | 0.5 | Up to 5 Gigabit |
| t3.small | 24 |
576 | 2 | 20% | 1.37 | Up to 5 Gigabit |
| t3.medium | 24 |
576 | 2 | 20% | 3.09 | Up to 5 Gigabit |
| t2.micro | 6 |
144 | 1 | 10% | 0.5 | Low to moderate |
| t2.small | 12 |
288 | 1 | 20% | 1.55 | Low to moderate |
| t2.medium | 24 |
576 | 2 | 20% | 3.22 | Low to moderate |
* The number of credits that can be accrued is equivalent to the number of credits that can be earned in a 24-hour period.
** The baseline performance in the table is per vCPU. Some node sizes that have more than one vCPU. For these, calculate the baseline CPU utilization for the node by multiplying the vCPU percentage by the number of vCPUs.
The following CPU credit metrics are available for T3 and T4g burstable performance instances:
Note
These metrics are not available for T2 burstable performance instances.
-
CPUCreditUsage -
CPUCreditBalance
For more information on these metrics, see CPU Credit Metrics
In addition, be aware of these details:
-
All current generation node types are created in a virtual private cloud (VPC) based on Amazon VPC by default.
-
Redis OSS append-only files (AOF) aren't supported for T2 instances. Redis OSS configuration variables
appendonlyandappendfsyncaren't supported on Redis OSS version 2.8.22 and later.