摘要 |
This server for distributing messages quickly detects a processing-performance decline in a downstream server by using a method for keeping, for each distribution-destination server, a threshold such as a connection number and a time interval separate from a response timeout, and determining that processing performance has declined when the threshold is exceeded. In addition, the present invention improves system availability by identifying a server group in which processing performance has declined on the basis of a correlation pertaining to inter-server cooperative processing, when a distribution server exhibits a decline in processing performance, and by the distribution server distributing a message to a server not exhibiting a decline in processing performance. |