etcd

• Written in Go
• A distributed key value store
• provides a reliable way to store data across a cluster of machines.
• handles leader elections during network partitions and will tolerate machine failure, including the leader.
• Communication between etcd machines is handled via the Raft consensus algorithm.

Ref:

• https://www.youtube.com/watch?v=hQigKX0MxPw
• https://godoc.org/github.com/coreos/etcd/raft
• http://www.infoq.com/cn/articles/coreos-analyse-etcd
• https://medium.com/jorgeacetozi/kubernetes-master-components-etcd-api-server-controller-manager-and-scheduler-3a0179fc8186

Raft: The Understandable Distributed Consensus Protocol

Ref:

• https://speakerdeck.com/benbjohnson/raft-the-understandable-distributed-consensus-protocol/
• https://raft.github.io/

Raft

• A consensus algorithm
• Defines three different roles (Leader, Follower, and Candidate) and achieves consensus via an elected leader.

what is consensus

• Consensus involves multiple servers agreeing on values.
• Typical consensus algorithms make progress when any majority of their servers is available
• Eg. A cluster of 5 servers can continue to operate even if 2 servers fail. If more servers fail, they stop making progress (but will never return an incorrect result).
• A general approach to building fault-tolerant systems.
• Each server has a state machine and a log
  • state machine
    • fault-tolerant component
    • Each state machine takes as input commands from its log