White Paper: How To Resolve Network Partitions In Seconds With Lightbend Enterprise Suite
Even though Reactive systems can self-heal, the same is not true for network partitions, a.k.a. “split brain” scenarios...
Highly-distributed, real-time applications, regularly suffer network partitions, which result in one or more nodes in a cluster suddenly, and mysteriously, becoming unresponsive. It’s not a question of if, but rather when, a network partition will occur in your distributed, clustered systems.
In this technical white paper, we review the tricky, unavoidable nature of network partitions, and describe what kind of strategies are needed resolve split brain scenarios intelligently, in just seconds.
Our Split Brain Resolver (SBR) feature for Akka Cluster (part of Lightbend Reactive Platform's commercial Enterprise Suite) offers four such strategies–called “Keep Majority”, “Static Quorum”, “Keep Oldest”, and “Keep Referee”–that automatically execute in a matter of seconds if and when network partitions occur.