Zab protocol

E705296

The Zab protocol is a crash-recovery atomic broadcast and consensus algorithm used by Apache ZooKeeper to ensure ordered, reliable updates across distributed nodes.

Try in SPARQL Jump to: Statements Referenced by

Statements (45)

Predicate Object
instanceOf atomic broadcast protocol
consensus algorithm
crash-recovery protocol
assumes non-Byzantine failures
reliable communication channels
designedFor distributed systems
replicated state machines
ensures committed transactions survive leader changes
followers apply transactions in the same order as leader
only one active leader at a time
ensuresProperty atomic broadcast
durable updates
ordered updates
reliable broadcast
total order broadcast
fullName ZooKeeper Atomic Broadcast protocol NERFINISHED
guarantees agreement among non-faulty nodes
liveness under partial failures
no data loss for committed transactions
prefix ordering of committed transactions
handles follower crash recovery
leader crash recovery
hasAbbreviation Zab NERFINISHED
hasPhase broadcast phase
discovery phase
leader election phase
synchronization phase
hasProperty leader-based atomic broadcast
log-based state replication
primary-backup style replication
implementedIn Apache ZooKeeper server codebase
inspiredBy Viewstamped Replication NERFINISHED
maintains global order of ZooKeeper transactions
publication “Zab: High-performance broadcast for primary-backup systems” NERFINISHED
relatedTo Paxos NERFINISHED
Raft consensus algorithm NERFINISHED
requires distinguished leader
supportsModel crash-recovery failure model
usedBy Apache Hadoop ecosystem components via ZooKeeper
Apache ZooKeeper NERFINISHED
usedFor coordination in distributed applications
leader-based replication
replicated log management
usedIn ZooKeeper quorum-based write operations
ZooKeeper transaction ordering

Referenced by (1)

Full triples — surface form annotated when it differs from this entity's canonical label.

Apache ZooKeeper consensusAlgorithm Zab protocol