Redis Cluster Mode

All write requests are sent to the master node, which then synchronizes data to the slave nodes. The master node is primarily responsible for executing write operations and data synchronization, while the slave nodes handle read operations and alleviate system read pressure.

Sentinel Mode

Building upon the master-slave architecture, a sentinel role adds monitoring capabilities for cluster health status. Sentinels achieve this by sending commands to Redis servers to retrieve their operational state.

Sentinel Election

The election mechanism used by sentinels is based on the Raft consensus algorithm.

Sharding Mode

Redis Cluster implements data sharding and replication across multiple Redis nodes.

• Leveraging the automatic sharding capabilities of Redis Cluster, we can easily perform horizontal scaling to increase cluster throughput.
• Utilizing the built-in replication features of Redis Cluster ensures that if a subset of nodes fails or becomes unreachable, the cluster remains available by serving requests based on replica data, thereby enhancing overall availability.