A Go implementation of a distributed in-memory cache system with load balancing and consistency guarantees for read operations. The system implements a distributed architecture where multiple cache nodes work together to provide scalable and reliable caching services.
- Distributed caching across multiple nodes
- Load balancing for distributing requests
- Consistent reads during synchronization
- TTL (Time-To-Live) support for cache entries
- Automatic synchronization between nodes
- Round-robin load balancing
- HTTP API interface
- Clone the repository
- Ensure Go is installed on your system
go run main.go -mode=load_balancer -nodes=localhost:8081,localhost:8082,localhost:8083
go run main.go -mode=node -port=8081 -nodes=localhost:8081,localhost:8082,localhost:8083
go run main.go -mode=node -port=8082 -nodes=localhost:8081,localhost:8082,localhost:8083
go run main.go -mode=node -port=8083 -nodes=localhost:8081,localhost:8082,localhost:8083
POST /cache/set?key=mykey&value=myvalue&ttl=3600GET /cache/get?key=mykey-mode: Operating mode ("node" or "load_balancer")-port: Port number for the server-nodes: Comma-separated list of cache node addresses
- Load Balancer Port: 8080
- Default Node Ports: 8081, 8082, 8083
- Default Node List: localhost:8081,localhost:8082,localhost:8083
- Stores key-value pairs in memory
- Handles TTL expiration
- Maintains synchronization with other nodes
- Tracks last writer information for consistency
- Distributes requests across cache nodes
- Uses round-robin algorithm
- Provides single entry point for clients
The system ensures consistency through:
- Write synchronization across nodes
- Last writer tracking
- Read redirects during synchronization
- Atomic operations with mutex locks
- In-memory storage (data is lost on restart)
- Basic round-robin load balancing
- No persistent storage
- No automatic node discovery
- No failover handling