|
This commit adds a complete implementation of the Raft consensus algorithm,
providing a robust solution for achieving consensus in distributed systems.
Key features implemented:
- Three-state system: Follower, Candidate, and Leader states
- Leader election with randomized election timeouts (150-300ms)
- Log replication for state machine commands
- Heartbeat mechanism to maintain leader authority
- Safety guarantees through term numbers and log consistency checks
- Proper handling of split votes and network partitions
Implementation details:
- Added VSRaftProtocol.java with full Raft algorithm logic
- Integrated with existing event-driven simulation framework
- Supports dynamic cluster sizes with proper quorum calculations
- Implements RequestVote and AppendEntries RPCs
- Maintains persistent state (currentTerm, votedFor, log entries)
- Includes comprehensive logging for debugging and visualization
Testing:
- Added VSRaftProtocolTest.java with unit tests covering:
- Leader election scenarios
- Log replication mechanics
- State transitions
- Message handling for all RPC types
Integration:
- Registered protocol in VSRegisteredEvents for simulator discovery
- Added human-readable names in VSDefaultPrefs for UI display
- Compatible with existing visualization and timing systems
This implementation follows the Raft paper closely while adapting to
the simulator's event-driven architecture and message-passing model.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
|
