A study of service reliability and availability for distributed systems

Distributed systems are usually designed and developed to provide certain important services such as in computing and communication systems. In this paper, a general model is presented for a centralized heterogeneous distributed system, which is widely used in distributed system design. Based on this model, the distributed service reliability which is defined as the probability of successfully providing the service in a distributed environment, an important performance measure for this type of systems, is investigated. An application example is used to illustrate the procedure. Furthermore, with the help of the model, various issues such as the release time to achieve a service reliability requirement, and the sensitivity of model parameters are studied. This type of analysis is important in the application of this type of models. © 2002 Elsevier Science Ltd. All rights reserved.