High service availability in MaTRICS for the OCS

Abstract

Internet services are becoming an integral part of our daily environment. Nobody sees the background of the configuration of the services and its backend systems, but it is expected that critical services, like flight booking systems, online banking, etc. are resilient against server and service faults. Therefore those services often run in a cluster on several machines in an Active/Standby configuration in a clustered mode. We present how we realize remote configuration and fault tolerance of the Online Conference Service (OCS) with our service oriented framework MaTRICS. Our solution lets the services untouched and uses the cluster management software heartbeat to provide high availability. © 2008 Springer-Verlag.