The Open Service Availability Framework (OpenSAF) is designed to work in a loosely coupled computer system with fast interconnections between the nodes. An application using OpenSAF could provide “five nines” (99.999 %) availability without re-inventing availability techniques. It provides mechanisms to replicate data between nodes, so that in case of failure the services can be started on a different node with the same state. It monitors itself for hardware and software failures, can keep several service instances running (with different availability models), and can automatically restart and migrate services to satisfy the availability requirements. It is based on the standard defined at saforum.org. It includes an availability management framework, a message service, a lock service, an event service, a checkpoint service, and much more.
