Performance-Related Reliability for Fault Tolerant Computing Systems

For computing systems in which there is no distinct difference between success and failure states, it is difficult to properly evaluate their fault-tolerant characteristics. For this reason, many pelformance-related reliability measures have been proposed. In this paper these measures, classified into three categories （i. e. measures based on instantaneous performances, measures based on cumulated performances, and profust reliability measures）, are introduced. Their solution methods are also briefly given here.