Abstract: | The reliability of a distributed program is very sensitive to the ways of file distribution, and communication system topology. It is usually much more complex to analyze the reliability of distributed program than that of the network reliability. This paper presents a distributed program reliability analysis algorithm that is developed based on the concept of sharp operation. The algorithm provides a (E−n+2) * E memory space requirement and O(n+E+E*Nt) computation time, where E is the number of edges, n is the number of nodes, and Nt is the number of file spanning trees in the corresponding graph of distributed system, respectively. Compared with existing algorithms, the proposed algorithm is considered more economic both in time and space. |