DCS Redis cluster instances are directly connected and compatible with the open-source Redis Cluster. Based on a decentralized cluster deployment architecture, each node in the cluster stores part of the data.
Features of Redis cluster instances:
l The smart client JedisCluster is supported.
l Compared with stand-alone instances, the overall performance is almost linearly proportional to the number of Redis shards.
Architecture Diagram
In the decentralized architecture of the Redis cluster, data storage and load processing tasks are not managed by a single central node, but are shared by multiple nodes. This architectural design aims to improve the scalability, availability and fault tolerance of the system. Data is distributed across multiple nodes, allowing for better performance and reliability.
Data Sharding
In a cluster, data sharding divides the entire data set into multiple fragments and stores them on different nodes. This sharding mechanism brings many benefits, including better horizontal scaling, load balancing, and system performance. A cluster is pre-configured with 16384 slots, and each Redis server stores the mapping relationships between all slots and the Redis server.
Application Scenarios
l Large Data Volume
Supports a single Redis shard specification of 1-64G, and the number of Redis shards can be selected from 3-64G. Supports a maximum specification of 4TB, which can effectively meet business expansion requirements.
l Scenarios with High QPS Pressure
Multi-node deployment is adopted to break the performance bottleneck of single-thread Redis, and to better support scenarios with high QPS.
l Throughput-Intensive Application Scenarios
Provides high performance, HA and horizontal scaling, suitable for scenarios with large numbers of read-write requests.
l Scenarios with High Redis Protocol Compatibility Requirements
Compatible with native clusters from the Redis community, and each business can be smoothly migrated.
l Scenarios without HA of Cached Data
Since the single-node cluster instance adopts a single-replica cluster deployment architecture, it does not have HA and service continuity. Please be sure to note the risks if you use it.