Skip to content

Redis Architecture and Scenarios

house edited this page Nov 22, 2021 · 1 revision

Redis Architecture and Scenarios

Redis is currently a mainstream caching product due to its high performance and simple deployment, and due to its simple architecture implementation and deployment, it is very suitable for combining cloud native technology to achieve cloud autonomy. The following is an introduction to Redis and applicable scenarios

Primary and standby mode

td-redis-operator

Architecture:

  • As shown above, the active/standby architecture introduces sentinel as the HA component to achieve failover based on the synchronization of the active/standby instance.
  • The client watches the new configuration by monitoring the sentinel's events.

Scenes:

  • The main-standby architecture generally cannot store large amounts of memory.
  • Suitable for high frequency query of a small amount of hot data, QPS usually upper limit is 10w.

Cluster mode

td-redis-operator

Architecture:

  • Decentralized architecture design, data is distributed through hash slots.
  • The nodes communicate through the gossip protocol and participate in failover at the same time.

Scenes:

  • Usually, the scene is stored in the internal memory within 1T.
  • In order to avoid cross-node data requests, some simple query application scenarios are recommended.
Clone this wiki locally