Aeron Cluster provides an implementation of Raft Consensus upon which highly available services can be built.
Aeron Cluster has the following capabilities:
- sequences multiple client connections into a single, replicated log
- provides efficient fault tolerance with 2 or more nodes
- provides Aeron Archive endpoints for writing to snapshots and reading from snapshots on start
- allows 1 or more clustered services to be run, with support for inter-service sequenced messaging
- reliable, sequenced timers
- very high levels of performance, in particular with premium features such as kernel bypass when using the C media driver
Typical use cases include:
- host one or more central limit order books
- host RFQ, IOI or similar negotiation logic
- message sequencer
- game state for multiplayer games
- any other scenario where you require fault-tolerant, high performance state
Aeron Cluster is typically deployed on bare metal and cloud services. Adaptive have open sourced an Aeron Cluster sample that builds a simple application with a 3 node deployment on Docker and Kubernetes.