NOTE
3.4 Distributed Consensus Algorithm: Gossip
1. What Is Gossip - An algorithm proposed by Xerox for replicating data among multiple nodes in a distributed database. - Nodes continuously exchange information, and after a period of time all nodes in the cluster will know the complete information. 2. Why Gossip Is Needed 3. Gossip Algorithm Process - Each node periodically and randomly selects a connected node to spread messages.
This is a historical learning note and may contain outdated or incomplete understanding.
1. What Is Gossip?
- An algorithm proposed by Xerox for replicating data among multiple nodes in a distributed database.
- Nodes continuously exchange information, and after a period of time all nodes in the cluster will know the complete information.
2. Why Is Gossip Needed?
3. Gossip Algorithm Process
- Each node periodically and randomly selects a connected node to spread messages.
- After other nodes receive messages they have not seen before, they spread those messages to other connected nodes.
4. Problems with Gossip
4.1. Message Redundancy
- Because nodes are selected randomly, the same message may be sent to the same node multiple times.
4.2. Consistency Problems
- Messages spread to the entire network little by little, so consistency problems necessarily exist during this period.
Discussion
Sign in with GitHub to comment. Discussions are stored as GitHub Issues.View on GitHub