Reliability-metric varieties and their relationships

Although reliability theory is based on the concept of failures, the definition of a failure may be very different from different perspectives. To a hardware engineer, a failure means a part replacement and verification of the failure. To a manufacturing repair depot, a failure is a returned part. To the finance department, a failure is a warranty claim. To a service organization, a failure is a service call for corrective maintenance (as opposed to planned or preventive maintenance). To a customer, a failure is a degradation of service or capability of the product. These various definitions of failure lead to various types of reliability metrics, such as part replacement rate or service call rate, in addition to the classic reliability metrics such as constant failure rate. This paper describes many types of reliability metrics and their interrelationships, both in theory and in practice. For the practical metric relationships, we draw from our experience with many reliability metrics at the Tandem Business Unit of the Compaq Computer Corporation. Our main conclusions are: the classical engineering view of failures is not sufficient for customer satisfaction with reliability in our market; multiple reliability metrics need to be tracked to satisfy the needs of various organizations; and many different functional groups must work together to ensure customer satisfaction with reliability