Tendis is a distributed key-value store that implements the Redis protocol while using RocksDB as its storage engine.
Tendis addresses the limitation of Redis's in-memory storage by persisting all data to disk through RocksDB, enabling storage capacity far exceeding available memory. This approach allows applications to use Redis clients without modification while supporting petabyte-scale datasets. The system employs a decentralized distributed architecture similar to Redis Cluster, using the gossip protocol for inter-node communication and supporting automatic node discovery and failover when master nodes fail.
Tendis suits workloads where storage cost is a primary concern, particularly scenarios with significant cold data that would be expensive to keep in memory. The tool is designed as a hybrid storage component that can work alongside Redis to balance performance and cost, potentially reducing storage expenses substantially. Organizations should choose Tendis when they need Redis protocol compatibility but can tolerate disk-based latency in exchange for larger capacity and lower infrastructure costs. The system supports horizontal scaling to approximately one thousand nodes with linear performance characteristics and includes online data migration between cluster nodes.
The project maintains active development with regular commits addressing both feature additions and stability improvements. The codebase shows consistent refinement of the distributed coordination mechanisms and storage layer optimizations. Documentation and testing infrastructure receive ongoing attention, with test scripts included to validate changes. The project includes performance benchmarking data and maintains support channels for user questions and issue reporting.
