Markov theory based multi-server C/S system reliability algorithm

This paper applies Markov theory to multi-server C/S system reliability assessment. It simulates the state of storage failure by a fault server and a fault line, and regards it as communication line failure. When dumping of data fails also, transferring of data switches to using this fault communication line. It forms Markov transition probability matrix by utilizing connection relationship between servers and lines, fully considering the rate of dumping success when one line fails, and synthesizing components failure rate. Utilizing classic Markov state space method, it thoroughly, accurately and easily resolves C/S system reliability rate. Finally, it testifies the applicability of the algorithm by an actual C/S system example.