Similartity Network For Semantic Web Services Substitution
This addresses the challenge of handling failures and improving performance in Web services composition, but it is incremental as it builds on existing classification methods.
The paper tackles the problem of automating Web services substitution by proposing a network-based approach using similarity measures between operations, and results show it enables more detailed analysis of substitutable services on a benchmark.
Web services substitution is one of the most challenging tasks for automating the composition process of multiple Web services. It aims to improve performances and to deal efficiently with Web services failures. Many existing solutions have approached the problem through classification of substitutable Web services. To go a step further, we propose in this paper a network based approach where nodes are Web services operations and links join similar operations. Four similarity measures based on the comparison of input and output parameters values of Web services operations are presented. A comparative evaluation of the topological structure of the corresponding networks is performed on a benchmark of semantically annotated Web services. Results show that this approach allows a more detailed analysis of substitutable Web services.