A Rapid Algorithm to Find Replacement Services for K-Shortest Path Problem with QoS Constraints

During the execution of composite service, if one service component in critical path fails, or becomes overloaded not to be accessed, a mechanism is needed to ensure that the whole running process isn't interrupted. The failed service needs be replaced quickly and efficiently by switching to a new path. Recent researches on this problem don't consider the influence of the QoS criteria of service candidates or the connection state of the overlay network, so the algorithms are easily disabled. In this paper, we solve the maximum utility path problem that is motivated by recent interest in k-shortest path problem. A new algorithm WSC_KUPB (web service composition k-maximum utility path backup) is described to enumerate the k maximum utility of simple loopless paths in a digraph. We construct utility path tree by using the utility function as the evaluation standard. And we enumerate every edge in cut edge set across two utility path trees; compute the utility value of the new path. Finally the optimal path with the maximum total utility is chosen as a backup path. The test result shows the algorithm performs very well in the efficiency and availability as compared to other relevant algorithms.