摘要 |
PROBLEM TO BE SOLVED: To allow system setting in consideration of both processing performance in an active system and reliability securement of a cluster system using a volatile memory to solve the following problems: processing of the active system is delayed when a temporal increase in network and a congestion are generated; it requires time to confirm transfer of update data to part of standby systems; and when transfer of the update data to the standby system, which cannot be confirmed, is given up, reliability of the cluster system is deteriorated in the cluster system. SOLUTION: The active system performs multicast transmission of the update data to a plurality of standby systems. When responses are not returned, the system switches unicast transmission using a standby system from which the response is not returned as a destination; when the response is not returned within specified time, the system gives up synchronization of the standby system, and continues the following transactions. A ratio that response time of each standby system at fixed time intervals occupies in synchronization specified time, and statistical information until the standby system is synchronized after once giving up the synchronization are acquired; and information for reviewing specification of the optimal synchronization specified time, network configuration and applications of the cluster system is provided. COPYRIGHT: (C)2009,JPO&INPIT
|