A method for providing an application cluster service (APCS) with fault-detection and failure-recovery capabilities. This method is composed of the steps of nodes clustering, invoking and detecting applications, fault-recovery of applications, detection of nodes, and node replacement. This method is applicable in a clustered environment to detect if a slave node is failed by sending a heartbeat periodically from a master node; and to detect if the master node still exists by checking if the master node stops sending the heartbeat (i.e. the master node may be failed).
|Translated title of the contribution||提供具容錯能力之應用程式叢集服務的方法|
|Publication status||Published - 2005 Oct 27|