distributed-consensus-and-replication-pr-module

consensusAgreement on a single value among a set of distributed processes; the core problem consensus protocols solve.

quorumSubset of nodes sufficient for reads or writes; typically majority (n/2+1) for strong consistency.

replicated logOrdered log of commands maintained across replicas; foundation of state machine replication.

raftConsensus algorithm designed for understandability, using leader election, log replication, and safety viaterm

eventual consistencyGiven no new updates, all replicas will eventually converge to the same value.

a005Uniform agreement: all processes including those that later fail must decide on the same value, stricter than

failure detectorMechanism to distinguish crashed nodes from slow ones based on timeouts, heartbeats, or phi_accrual.

leader electionProcess of selecting a unique leader from among candidates when no leader exists.

replica synchronizationPropagation of updates to ensure all replicas converge toward identical state.

split brainScenario where network partition causes multiple nodes to believe they are leader; resolved by quorum.

a002Validity in agreement: the decided value must have been proposed by some process in the system, guaranteeing a

a007Quorum in agreement: a subset of processes sufficient to guarantee agreement can be reached, typically interse

a008Threshold in agreement: the minimum number of correct processes required to guarantee safety and liveness unde

a009Asymmetric agreement: one class of processes (e.g., primary) must agree while another class (e.g., backups) ma

agreement decideA value reaches agreement when a quorum of nodes have accepted it; once decided, the value is stable and no co

byzantine failureComponent exhibits arbitrary behavior including lies, omissions, or timing violations.

client sessionTracking mechanism associating client requests and ensuring exactly-once execution.

commit indexIndex of highest log entry known committed; entries up to this point are durable.

failure detectionMechanism for detecting crashed or unreachable nodes in a distributed system.

failure modeClassification of ways components can fail: crash_stop, crash_recovery, omission, timing, response, arbitrary_