Distribute Redis Architecture

Distributed Redis uses a Redis shard + Proxy architecture. Redis sharding is based on the master-replica Redis resource pool, easily breaking through the single-threaded bottleneck of Redis itself. It supports online expansion and can greatly meet the business requirements for large capacity or high performance of Redis.

Distributed Redis Architecture - Redis Sharding

The distributed version of Redis provides a default access IP. Users can access this IP for normal Redis access and data operations.

Redis Sharding: Each shard server is a master-replica version of Redis with a high-availability architecture. If the master node fails, the system will automatically switch between the master and the replica to ensure high availability of the service. The minimum capacity of a single shard is 1G. By default, an alarm will be set for the memory usage rate of a single shard. When the memory usage rate of a shard reaches the alarm threshold, an alarm will be triggered. In a real production environment, if the memory usage of a single shard is full, some data will not be able to be written. Therefore, if a shard has a high load or high memory usage, you can expand this shard separately. In addition, it is recommended that users try to balance the load among all shards to avoid a situation where one shard is very idle, causing resource waste, while another shard is heavily loaded.

Distributed Redis Architecture - Proxy

Proxy Proxy: Dual master node configuration, there will be multiple Proxies in the distributed architecture, and the system will automatically implement load balancing and failover for them.