BiteSizedChunks.comLearn one small thing at a time.

Course contentsShow

599What Is Distributed Consensus?
600Why Consensus Is Hard
601The FLP Impossibility Result
602Crash-Stop vs Byzantine Failures
603Consensus Use Cases
604Safety vs Liveness Properties
605Quorums and Majority Agreement
606The CAP Trade-off in Consensus
607Consensus vs Consistency Models
608The Problem Paxos Solves
609Paxos Safety and Liveness Guarantees
610The Three Roles in Paxos
611Proposal Numbers and Ordering
612The Two-Phase Protocol
613The Prepare Phase
614The Accept Phase
615Handling Conflicts and Preemption
616Multi-Paxos for Log Replication
617Why Paxos Is Difficult in Practice
618Raft Overview: Understandability as a Design Goal
619Server States: Leader, Follower, Candidate
620Terms: Logical Time in Raft
621Leader Election: RequestVote RPC
622Election Timeouts and Randomization
623Log Structure and Entries
624AppendEntries RPC: Replication Mechanism
625Log Matching Property
626Commitment and the Commit Index
627Safety: Leader Completeness Property
628Election Restriction: Up-to-Date Check
629Log Inconsistencies and Repair
630Safety Argument: Committing Entries from Current Term
631Membership Changes: Joint Consensus
632Log Compaction: Snapshotting
633ZooKeeper: Coordination Service Built on Consensus
634etcd: Distributed Key-Value Store with Raft
635Consul: Service Discovery with Raft Consensus
636Consensus for Leader Election
637Distributed Locks via Consensus
638Configuration Management with Consensus
639Consensus Cluster Sizing Tradeoffs
640Performance Characteristics of Consensus
641Consensus in Multi-Region Deployments
642When NOT to Use Consensus
643Monitoring and Operating Consensus Clusters

599What Is Distributed Consensus?
600Why Consensus Is Hard
601The FLP Impossibility Result
602Crash-Stop vs Byzantine Failures
603Consensus Use Cases
604Safety vs Liveness Properties
605Quorums and Majority Agreement
606The CAP Trade-off in Consensus
607Consensus vs Consistency Models
608The Problem Paxos Solves
609Paxos Safety and Liveness Guarantees
610The Three Roles in Paxos
611Proposal Numbers and Ordering
612The Two-Phase Protocol
613The Prepare Phase
614The Accept Phase
615Handling Conflicts and Preemption
616Multi-Paxos for Log Replication
617Why Paxos Is Difficult in Practice
618Raft Overview: Understandability as a Design Goal
619Server States: Leader, Follower, Candidate
620Terms: Logical Time in Raft
621Leader Election: RequestVote RPC
622Election Timeouts and Randomization
623Log Structure and Entries
624AppendEntries RPC: Replication Mechanism
625Log Matching Property
626Commitment and the Commit Index
627Safety: Leader Completeness Property
628Election Restriction: Up-to-Date Check
629Log Inconsistencies and Repair
630Safety Argument: Committing Entries from Current Term
631Membership Changes: Joint Consensus
632Log Compaction: Snapshotting
633ZooKeeper: Coordination Service Built on Consensus
634etcd: Distributed Key-Value Store with Raft
635Consul: Service Discovery with Raft Consensus
636Consensus for Leader Election
637Distributed Locks via Consensus
638Configuration Management with Consensus
639Consensus Cluster Sizing Tradeoffs
640Performance Characteristics of Consensus
641Consensus in Multi-Region Deployments
642When NOT to Use Consensus
643Monitoring and Operating Consensus Clusters

← System Design

Lesson 602 of 1,919·17. Consensus AlgorithmsPro lesson

Crash-Stop vs Byzantine Failures

Distinguish between nodes that simply stop versus nodes that behave arbitrarily or maliciously.

This lesson is for subscribers

You've completed the free preview. Subscribe to unlock every lesson in every course.

See pricing Back to course