TSINGHUA SCIENCE AND TECHNOLOG Efficient Composition of Semantic Web Service with End-to-End QoS Optimization

The Efficiency of QoS-aware service composition is important since most of the service composition problems are known to be NP-hard. With the growing number of Web services, the service composition is like a decision problem on selecting services or/and execution plan to satisfy user’s end-to-end QoS requirement (e.g. response time, throughput). Composite services with same functionality may differ in execution plan, which may cause different end-to-end QoS. We present a powerful conceptual model combining semantic data-link and QoS, which leads to an efficient approach to automatically constructing a composite service with optimal end-to-end QoS. Our approach is based on greedy algorithm to select both services and execution plan for composite service. We give empirical and theoretical analyses of the proposed approach and show its time complexity of O(mn), for a repository with n services and an ontology with m concepts. However, the approach takes linear time in practice when we use index to search services in repository. Repository with 20,000 services and ontology with 300,000 concepts at most are used in the experiments, and the evaluations show that our algorithm significantly outperforms current existing algorithms in terms of composition efficiency while achieving 100% optimal end-to-end QoS.