Implementing operational profiles to measure system reliability

A method is presented, devised for creating operational profiles (OPs) by `decomposing´ an OP into two components: a configuration profile (CP) and a usage profile (UP). This decomposition is necessary in order to reduce the broad variation in customer usage into a set of test environments (OPs) that can be implemented as part of a commercial product development/test process. These test environments must, of course, emulate actual customers closely enough so that the data gathered from the testing can be used to predict the reliability observed in the field. The author describes the decomposition of an operational profile into its components and provides examples from the project in which this method was first used (an AT&T International Private Branch Exchange development)