This page describes the differences between the available Redis tiers.

Memorystore for Redis supports Basic and Standard
Tiers. The Basic Tier is best suited for applications that use Redis as a cache
and can withstand a cold restart and full cache flush. Standard Tier instances
provide [High Availability](https://docs.cloud.google.com/memorystore/docs/redis/high-availability)
using replication, read replicas, and automatic failover.

To help protect your data and provide a baseline level of persistence, RDB
snapshots are enabled by default when you create a Memorystore for Redis
instance using the Google Cloud console.

It is not possible to upgrade from Basic Tier to Standard Tier or the other way
around. If you want to move data from an existing Basic Tier instance to a new
Standard Tier instance, use [import and export](https://docs.cloud.google.com/memorystore/docs/redis/import-export-overview)
to move your data.

## Tier architecture

The following diagrams show the architectures for the
Memorystore for Redis service tiers.

### Basic Tier

![](https://docs.cloud.google.com/static/memorystore/img/basic-tier.png)

- One Redis node
- Ephemeral cache

### Standard Tier

![](https://docs.cloud.google.com/static/memorystore/img/standard-tier.png)

- Provides [High Availability](https://docs.cloud.google.com/memorystore/docs/redis/high-availability) with replication
- Fails over to replica in the case of primary node failure

### Standard Tier with read replicas

![](https://docs.cloud.google.com/static/memorystore/img/standard-tier-read-replicas.png)

- Provides [High Availability](https://docs.cloud.google.com/memorystore/docs/redis/high-availability) with replication
- Read replicas allow for distributed reads

## Feature comparison

| Specifications | Basic Tier | Standard Tier (read replicas disabled) | Standard Tier (read replicas enabled) |
|---|---|---|---|
| **Description** | Provides a cache with no replication | Provides redundancy and availability using replication | Provides redundancy and availability using replication to backup data, and multiple Read replicas to increase read throughput |
| **Max Redis primary size** | 300 GB | 300 GB | 300 GB |
| **Max network bandwidth** | 16 Gbps | 16 Gbps | 16 Gbps total for writes. 16 Gbps per node for reads.^1^ |
| **I/O threads** | Yes^2^ | Yes^2^ | Yes^2^ |
| **Scale primary size** | Yes | Yes | Yes |
| **Scale number of read replicas** | No | No | Yes |
| **Cross-zone replication** | No | Yes^3^ | Yes^3^ |
| **Automatic failover** | No | Yes | Yes |
| **Read replicas** | No | No | Yes |
| **In-transit encryption** | Yes | Yes | Yes |
| **Maintenance window** | Yes | Yes | Yes |
| **Cloud Monitoring** | Yes | Yes | Yes |

^1^ The maximum write throughput is 16 Gbps. Read throughput is dependent on the number of nodes (read replicas) in the instance, including the primary node. For example, if you have an instance with 1 primary node and 2 read replicas, the total read throughput is 48 Gbps.

^2^ I/O threads are only available on M3 instances or higher running Redis version 6.x. For more information, see [Redis version 6.x](https://docs.cloud.google.com/memorystore/docs/redis/supported-versions.md.txt#redis_version_6x).

^3^ For more information about region-specific considerations, see [Geography and regions](https://docs.cloud.google.com/docs/geography-and-regions.md.txt#regions_and_zones).