MemKV is a high-performance, in-memory key-value database built from scratch in Go. It is inspired by Redis and implements the RESP (Redis Serialization Protocol), making it fully compatible with redis-cli.

This project was built for educational purposes to explore high-performance network servers, custom data structures, and the low-level design of modern databases.

• Redis Protocol (RESP) Compliant: Fully compatible with redis-cli and other Redis clients.

• High-Performance I/O: Uses a single-threaded, event-loop architecture with I/O multiplexing (epoll for Linux and kqueue for macOS) to handle thousands of concurrent connections.

• Custom Data Structures: Implements complex data structures from scratch, including:

  • Skip List: For high-performance sorted sets (ZADD, ZRANK, etc.).
  • Geohash: For efficient geospatial indexing (GEOADD, GEODIST, etc.).

• Probabilistic Data Structures: Includes implementations of:

  • Scalable Bloom Filter: For fast, memory-efficient set membership testing (BF.ADD, BF.EXISTS).

  • Count-Min Sketch: For estimating item frequencies in a data stream (CMS.INCRBY, CMS.QUERY).

• Graceful Shutdown: Ensures data is handled correctly and connections are closed properly on server termination.

  cd cmd
  go run main.go
  # on another terminal
  redis-cli -p 8081

The server supports a wide range of commands grouped by data type:

[ ] Hyperloglog

[ ] Morris counter

[ ] Cuckoo filter

[ ] Approx LRU eviction

[ ] Approx LFU eviction

[ ] Longest Common Subsequence