Parallel Optimization for Data-Intensive Service Composition

Abstract

The age of Big Data has inspired the appearance and application of data-intensive Web service. In most cases, multiple data-intensive services are assembled into a service composition to meet complicated requirements. As the number of Data-intensive Web services on the Internet is increasing rapidly and dramatically, traditional central service composition approaches have come to a performance bottleneck. This paper proposes a method for automatic data-intensive service composition, which can be executed in parallel. Firstly, the problem of automatic service composition is defined by combing the approaches of State Space Search and Planning Graph. Then, a heuristic algorithm is proposed to compose data-intensive Web service automatically. After that, more details are given to present the parallel optimization for the composition algorithm. A series of experiments show that the proposed parallel optimization method improves the efficiency of automatic Web service composition to a great extent.