Cluster federation allows users to efficiently scale up their application deployment and management to a global scale. Global scheduling allows replication and failover to span across physical datacenters and cloud providers, yielding lower latencies and higher availability for end users.
Kubernetes 1.3 introduced support for cluster federation, which is a natural extension to the work happening in Kubernetes and is one of those features that everyone dreams of having but few organizations achieve. Painting the picture of unified global deployments and showing these features working will help to attract more developers and operators to the Kubernetes platform. Federation brings a level of automation to cluster management that greatly reduces the operational overhead
This talk will show how to distribute workloads easily across clusters, regardless of vendor or physical location. It will also describe potential use cases and walk through future work.