Estimating the reliability of distributed systems

Reliability is a desired, yet hard to achieve feature of a distributed system. This feature is hard to quantify due to several intricate relationships that exist among the benefits of a distributed system. A conceptual model along with the algorithms for computing the reliability of a distributed system and distributed task, based on graph-theoretic formalism are presented. The algorithms are the following: Distributed Task Reliability with Cost Constraint (DTRC), Distributed Task Reliability with Articulation Point Constraint (DTRA), Distributed System Reliability with Cost Constraint (DSRC), and Distributed System Reliability with Articulation Point Constraint (DSRA). The algorithms generate minimum file spanning trees which satisfy a certain constraint. The constraint can either be a cost or an articulation point that directly affects reliability of a distributed system. This paper presents a graph theoretic model that is used to come up with an estimate of a distributed system\u27s reliability. An analysis of the results is also presented to verify the correctness of the algorithms