Topic · topic
Consensus
Consensus is the distributed systems problem of getting a set of unreliable processes to agree on a value or sequence of values. Consensus algorithms power state-machine replication for databases, key-value stores, configuration systems, and blockchains. The consensus topic spans classical crash-fault tolerant algorithms (Paxos, Raft, Multi-Paxos, Zab, Viewstamped Replication) and Byzantine fault tolerant algorithms (PBFT, Tendermint, HotStuff, Casper FFG/CBC), and the impossibility result FLP that bounds what is possible in fully asynchronous models.
Resources
-
Paxos
Paxos is the classical crash-fault-tolerant consensus algorithm introduced by Leslie Lamport in 1989. Paxos and its derivatives (Multi-Paxos, Cheap Paxos, Vertical Paxos, EPaxos, Flexible Paxos) are foundational to distributed systems theo…
-
Raft
Raft is a consensus algorithm designed for understandability, published by Ongaro and Ousterhout in 2014. Raft separates leader election, log replication, and safety into distinct concepts and is implemented widely in etcd, Consul, Cockroa…
-
Practical Byzantine Fault Tolerance (PBFT)
PBFT, by Castro and Liskov (1999), is a Byzantine fault-tolerant consensus protocol that tolerates up to f Byzantine failures with 3f+1 replicas. PBFT influenced subsequent BFT protocols including Tendermint and HotStuff.
-
Tendermint / CometBFT
Tendermint (now CometBFT) is a Byzantine fault-tolerant consensus engine that powers Cosmos SDK chains. CometBFT decouples consensus from application logic via the Application Blockchain Interface (ABCI), allowing custom blockchain applica…
-
HotStuff
HotStuff is a leader-based BFT consensus protocol with linear view change and three-chain commit, designed by Yin et al. HotStuff is the foundation for Diem/Libra BFT and inspired modern protocols like Aptos AptosBFT and Sui's Mysticeti.
Links
IssueTrackerReleasesSecurityPolicyCodeOfConductContributionGuideLicenseDomainSecurityReferenceReferenceReferenceReferenceReferenceResources
Providers working in Consensus
Providers whose own tags share at least two of this topic's tags, most shared first.
| Provider | About | Rating | APIs |
|---|---|---|---|
| Raft | Raft is a consensus algorithm for distributed systems designed to be understandable and provide the same guarantees as Paxos, used for distributed log replication and leader election in fault-tolerant systems. | minimal | 0 |
| Etcd | etcd is a CNCF graduated distributed, reliable key-value store used as the backing store for all Kubernetes cluster data. It provides strong consistency guarantees using the Raft consensus algorithm, supporting watch operations, lease-base… | developing | 1 |
| Apache Helix | Apache Helix is a generic cluster management framework for partitioned and replicated distributed resources. It automates partition management, replication, fault tolerance, and cluster expansion for distributed systems, providing a REST A… | developing | 1 |
| Sapien | Sapien is the company behind Proof of Quality (PoQ), an open protocol and consensus/attestation system for verifiable quality signals on AI data and subjective expert outputs. A panel of independent, collateral-backed validators reviews ea… | developing | 1 |
| Tendermint | Tendermint is a core contributor to the Cosmos Network and the original developer of Tendermint Core, a best-in-class Byzantine Fault Tolerant (BFT) consensus engine for state-machine replication, alongside the Cosmos SDK blockchain applic… | thin | 1 |
| Espresso | Espresso Systems builds the Espresso Network, a high-performance consensus and sequencing layer that gives rollups and institution-grade financial applications real-time settlement (~3 second finality) without sacrificing control, privacy,… | thin | 1 |
| Cypherium | Cypherium is a permissionless Layer-1 blockchain built to bridge centralized (CeFi) and decentralized (DeFi) finance and bring real-world assets on-chain at scale. It runs CypherBFT, a hybrid consensus that pairs GPU proof-of-work committe… | emerging | 2 |
Tags
AlgorithmsBFTBlockchainConsensusCrash Fault ToleranceDistributed SystemsReplicationState Machine