A backtracking approach for semantic matchmaking algorithm in Web service discovery

Abstract

Web service discovery in Service Oriented Architecture has been an important research field for optimization as the use of internet for this purpose has been drastically seen to increase over the years. There are lot many proposals on improving match making algorithm in Web service discovery process to result appropriate target web services. Semantic web service search techniques are continued to be adapted after analyzing the short comings of syntax based or keyword based matching. However, the search time for the web services in most algorithms are high so as to realize a high precision rate in the results. A new backtracking approach is proposed in this paper that generates a sorted list of available operation interface annotations which is used to backtrack to the services that are most matching with the query service. The method is implemented to test its precision and performance with Brute force, Paolucci's Greedy matchmaking and Assignment Algorithm and to analyze various parameters influencing the algorithm for understanding the suitable environments in which the algorithm can run better. Results are achieved in much better running time than other algorithms and give a better precision rate than Paolucci's matchmaking algorithm in very large test cases.