Open Cluster Framework
Welcome to the Open Cluster Framework project. The purpose of this project is to define standard clustering APIs for some basic capabilities.
We are a group of people who have interests in cluster software - both providers and consumers of clustering services. We have periodic meetings in person, and an ongoing conversations via a mailing list.
We have two basic thrusts to our work.
Create and support an open source development project which acts as the reference implementation for the OCF APIs.
The established open source reference implementation for High Availability Clusters is Pacemaker.
It is also our intent to create APIs which are usable on both High-Performance and High-Availability clusters. It is not our intent to replace or redefine de-facto standards (like MPI or PVM) which are already in common use, and serve their intended purpose well.
It is our stance that the OCF APIs themselves must be royalty-free (RF) standards. It is acceptable for individual implementations to use patented or otherwise encumbered techniques, but the standard itself must be reasonably implementable without the use of patented techniques.
We are a working group of the Free Standards Group. The standards themselves are intended to be largely platform-independent, capable of being implemented on most POSIX-compliant OSes, but there will be certain sections (kernel APIs for example) which will be platform-specific. Although the standards are intended to be OS-independent, the primary interest of the majority of the group is Linux, and the OS-specific APIs will likely only be defined for Linux initially.