Policy management for distributed computing and a method for aging statistics

摘要

A policy management system and method having a plurality of cooperating computers connected in a network. A policy management software resident in one or more managing computers of the network monitors the network and collects performance related values, such as, response time or queue delay of the cooperating computers. Performance related metrics are derived from the performance values and posted for access by the software that distributes work or controls execution of the work. The performance metrics includes only a number and average of the values received. The performance related values are formed in a data structure having n rows, where n is the number of value reporting intervals for which the performance metrics are kept. As a new current interval begins, the performance metrics of the nth row of a preceding interval are discarded and such nth row is used as the first row for the current interval. The remaining rows are shifted down one row position.