A composite web services discovery technique based on community mining

Abstract

Community structure has been recognized as an important statistical feature of network systems over the past decade. The web service in SOA system naturally forms into some service community during execution process, within which the links between nodes are very dense, but between which they are quite sparse. These service communities were generated by repeatedly interaction between composite services which accomplish the same task. Mining and analysis web service community will help design SOA system and predict service behavior. This article addresses the problem of how to discovering and quantifying web services community formed by closely interactive web services and gives the composite web service discovery technique. We consider the case where the details usage record is logging by execution engine. We proposed a novel approach which construct web service interactive network (WSIN) from usage log and get community structure by spectrum clustering. Generally, the web services belong to same cluster have strong relative to same task object and we call it web service community. The approach has been implemented in an experience system for web services dynamic composition and discovery, and the experimental results demonstrated the efficiency and effectiveness of the proposed algorithm.