Empirical Study on the Characteristics of Complex Networks in Mashup Service Network

With the rapid increase and wide application of Web APIs and Mashup services, the relationship among Web APIs, Mashup services and their tags become very complex and constantly evolve, result in forming a large network of services. And the characteristics research on this network by using complex network has an important significance, which will greatly promote service discovery and composition. In this paper, we conduct the empirical study to the large-scale Mashup service network. Firstly, Web APIs, Mashup services and tags are crawled from ProgrammableWeb and Mashup service network prototype system is developed, and Mashup service homogeneous and heterogeneous networks are built. Then, we use degree and degree distribution, average shortest path length and clustering coefficient of complex network, to measure Mashup service homogeneous and heterogeneous networks. The experimental results show that Mashup service network has the characteristics of small-world and scale-free, which will provide some guidance for service discovery and composition.