Gossip Protocol & Anti-Entropy
ReliabilityNodes periodically exchange state with a few random peers, spreading information (or repairing inconsistencies) across a whole cluster without any central coordinator.
Gossip protocols spread information the way rumors spread through a social network: each node periodically picks a few random peers and exchanges what it knows (cluster membership, node health, or data updates). Over a small number of rounds, information reaches the whole cluster with no central coordinator and no single point of failure — the cost of a node being temporarily unreachable is just slightly slower propagation, not a hard failure. Anti-entropy is the closely related idea of nodes comparing data (often via Merkle trees, so they only need to exchange hashes of hashes to find the actual differing keys) and repairing any inconsistencies found, used to converge replicas that drifted apart.
How it connects
Gossip Protocol & Anti-Entropy as the source, with the components it typically interacts with.
- → Consensus & Replication Protocols: Gossip is a weaker, more scalable alternative to full consensus for propagating cluster membership/state when strict agreement isn't required.
- → Database Types (SQL, NoSQL & Beyond): Leaderless databases (Cassandra, DynamoDB-style) use gossip to propagate node membership and health across the cluster without a central coordinator.
- → Vector Clocks & CRDTs: Gossiped updates in a leaderless system often carry vector clocks/CRDT metadata so nodes can merge concurrently gossiped state deterministically.