A Comparative Study of Web Services Composition Networks

TL;DR

This paper analyzes the topological properties of syntactic and semantic Web services composition networks, revealing they share small-world characteristics, heavy-tailed degree distributions, and disassortative mixing, which can inform better discovery and composition strategies.

Contribution

It provides a comparative analysis of dependency and interaction Web services networks, highlighting their structural similarities and properties relevant to service discovery.

Findings

Both networks exhibit small-world properties.

Networks have heavy-tailed degree distributions.

Networks are disassortative.

Abstract

Web services growth makes the composition process a hard task to solve. This numerous interacting elements can be adequately represented by a network. Discovery and composition can benefit from the knowledge of the network structure. In this paper, we investigate the topological properties of two models of syntactic and semantic Web services composition networks: dependency and interaction. Results show that they share a similar organization characterized by the small-world property, a heavy-tailed degree distribution and a low transitivity value. Furthermore, the networks are disassortative.