{"@context":"https://schema.org","@type":"CreativeWork","@id":"https://forgecascade.org/public/capsules/8cea273c-8543-438f-beb9-9bdd4ecc297e","name":"Distributed Systems: Consensus Algorithms and CAP Theorem","text":"CAP theorem: a distributed system can guarantee at most 2 of: Consistency, Availability, Partition tolerance. Paxos: leader election, prepare/promise/accept/commit phases. Raft: leader election by term, log replication, majority quorum. ZooKeeper: Apache coordination service, ZAB protocol. etcd: Raft-based key-value store used by Kubernetes. Byzantine fault tolerance: BFT requires 3f+1 nodes to tolerate f Byzantine failures. PBFT (Practical BFT): O(n²) message complexity. Tendermint: BFT consensus for blockchains. CRDT (Conflict-free replicated data types): G-counter, PN-counter, LWW register. Vector clocks: causality tracking. Lamport timestamps. Two-phase commit (2PC): coordinator + participants, blocking failure mode. Three-phase commit (3PC): adds pre-commit phase. Forge: distributed systems capsules linking CS theory and infrastructure.","keywords":["distributed-systems","consensus","cs-theory"],"about":[{"@type":"Thing","name":"SyncAppvPublishingServer"},{"@type":"Thing","name":"Forge Web Credentials"},{"@type":"Thing","name":"Startup Items"},{"@type":"Thing","name":"Unknown Logger"},{"@type":"Thing","name":"PsExec"},{"@type":"Thing","name":"ADVSTORESHELL"},{"@type":"Thing","name":"Group5"},{"@type":"Thing","name":"Axiom"},{"@type":"Thing","name":"APT30"}],"citation":[],"isPartOf":{"@type":"Dataset","name":"Forge Cascade Knowledge Graph","url":"https://forgecascade.org"},"publisher":{"@type":"Organization","name":"Forge Cascade","url":"https://forgecascade.org"}}