Exploiting web service usage context to facilitate services organization

Service-oriented architectures (SOA) are becoming the dominant computing paradigm, where all resources are abstracted as services to form services society. Open-access and easy-visiting of services are keys to services eco-system, thus service search engines will play an increasingly important role. Generally retrieved services are listed and ranked by the content similarities to queries. One problem is that if user clicks one service and finds that it is not good. Then how to find similar services to this service is a challenging problem. We find that traditional content-based similarity calculation method is insufficient to clustering similar services. In this paper, we introduce a concept of services "context" to characterize services. By using service context, we build services related collaboration graphs. A dynamic clustering approach is designed with respect to their contexts.