[cgl_discussion] USE CASE - Cluster Message Service (SA Forum AIS)
sdake at mvista.com
Mon Apr 18 15:29:50 PDT 2005
Here is the first run at the use case for cluster message service.
One technique for improving availability is to reduce the MTTR. This can
be achieved by distributing applications to mulitple hardware/operating
system instances. Then a service is replicated to multiple processors.
When any service fails, another standby service takes over.
After distributing an application in this fashion, these processors may
need to communicate with other processors. This could be achieved with
TCP/IP, or some other point to point protocol. However, the SA Forum
provides an abstraction, which once mastered, provides several
advantages to the developer.
A standard interface for increasing availability for Linux is
desireable. Unfortunately using TCP/IP with SA Forum application
communication is awkward. A SA Forum AIS messaging service removes the
awkward interface caused by intermixing AIS application failover and
* AIS Developers: Both open source and proprietary cluster
implementations are being developed with SA Forum interfaces.
* Application developers: Applications which require increased
availability by reducing MTTR are developed to the SA Forum AMF
interface specification. These applications communicate with the
messaging and eventing service as specified by SA Forum.
Applications and services that can benefit from using a the message
service are application failover services which require communicating
between redundant appications.
An application is developed which uses application failover. The
application desires to communicate with another application entirely. It
opens a communication channel and sends a request, then waits for a
No published messaging service exists which is B.01.01 compliant.
Linux-HA has a A.01.01 implementation available.
More information about the cgl_discussion