A Raft implementation inspired by XnnYygn.
I wanted to create something with the Netty framework and discovered the Raft consensus algorithm. Raft is a fascinating concept and, as the first distributed consensus algorithm I learned, I was motivated to implement its features after reading the original paper.
This project implements several key aspects of the Raft algorithm, including:
- Leader Election: Ensures that one server acts as the leader to manage the log replication process.
- Log Replication: Guarantees that logs are consistently replicated across all servers.
- Membership Change: Supports changes to the cluster membership (e.g., adding or removing a server).
- Log Compaction: Optimizes storage by compacting logs while maintaining the necessary state.
All of these features are encapsulated in the xraft-core module.
I believe that client interaction with the Raft implementation should be a service built on top of raft-core. To that end, I have created a simple key-value store called raft-state-machine. This store supports basic operations such as:
- GET: Retrieve the value associated with a specific key.
- SET: Store a value associated with a specific key.
To get started with Xraft, you can clone the repository and explore the modules. Detailed instructions on how to run the key-value store and interact with the Raft protocol will be provided in the documentation.