An Approach to Discover Dependencies between Service Operations

Abstract

Service composition is emerging as an important paradigm for constructing distributed applications by combining and reusing independently developed component services. One key issue of service composition is how to identify relevant service operations so as to compose services rapidly and correctly. A promising approach to simplifying the search of relevant service operations in service composition lies in the discovery of the dependencies between service operations. However, the discovery of operation dependencies is not a trivial task but a challenge. We propose an approach to discover operation dependencies in a personal problem solving environment. Our approach combines the semantic matching of inputs and outputs interfaces between service operations and the analysis of process cases to identify dependencies. The main contributions of the approach are: 1) It can be used to identify the direction of the dependencies. 2) It provides the method to measure the strength of dependencies. 3) Nonconflict property and non-redundancy property of discovered dependencies are guaranteed based on a dependency graph. Moreover, we experimentally demonstrate the efficacy of our approach by testing it under three typical bioinformatics scenarios.