Community Structure in Interaction Web Service Networks

TL;DR

This paper models interaction networks of Web services, revealing they share key complex network properties like small-world and scale-free features, and uncovers community structures with similar content and consistent topology.

Contribution

It introduces models of Web service interaction networks and provides an extensive analysis of their community structures and properties, highlighting their complex network characteristics.

Findings

Web service networks are small-world and scale-free.

Communities within these networks are similar in content.

Community structures are topologically consistent.

Abstract

Many real-world complex systems such as social, biological, information as well as technological systems results of a decentralized and unplanned evolution which leads to a common structuration. Irrespective of their origin, these so-called complex networks typically exhibit small-world and scale-free properties. Another common feature is their organisation into communities. In this paper, we introduce models of interaction networks based on the composition process of syntactic and semantic Web services. An extensive experimental study conducted on a benchmark of real Web services shows that these networks possess the typical properties of complex networks (small-world, scale-free). Unlike most social networks, they are not transitive. Using a representative sample of community detection algorithms, a community structuration is revealed. The comparative evaluation of the discovered community structures shows that they are very similar in terms of content. Furthermore, the analysis performed on the community structures and on the communities themselves, leads us to conclude that their topological properties are consistent.