Consensus Mechanisms
Overview
Implement distributed consensus among agent swarms. Select the appropriate protocol based on fault tolerance requirements, swarm size, and communication topology.
When to Use
- Multiple agents have produced independent solutions needing reconciliation
- Byzantine fault tolerance is required (untrusted or unreliable agents)
- State synchronization across distributed agent swarms
- Conflict-free data merging in concurrent operations
Protocols
| Protocol | Use Case | Fault Tolerance | Complexity | |----------|----------|-----------------|------------| | Raft | Leader-based consensus, ordered log | Crash faults (f < n/2) | Medium | | Byzantine | Untrusted agents, adversarial conditions | Byzantine faults (f < n/3) | High | | Gossip | Eventual consistency, state propagation | Partition tolerant | Low | | CRDT | Conflict-free replicated data types | Always convergent | Low |
Weighted Voting
- Queen agents: 3x weight multiplier
- Worker agents: 1x weight
- Configurable consensus threshold (majority, supermajority, unanimous)
Agents Used
agents/swarm-coordinator/- Protocol orchestrationagents/strategic-queen/- Weighted voting leadership
Tool Use
Invoke via babysitter process: methodologies/ruflo/ruflo-swarm-coordination