An Efficient Service Discovery Algorithm for Counting Bloom Filter-Based Service Registry

The Service registry, the yellow pages of Service-Oriented Architecture (SOA), plays a central role in SOA-based service systems. The service registry has to be scalable to manage large number of services along with their requirements on storage and discovery. Based on our previous work on feature-based services quantification, we characterize services according to their diverse functional and non-functional requirements, and represent them as string formats which can be stored, probed, and indexed by efficient data structures, such as hash table and Bloom filter. Then, we propose a comprehensive service-storage solution using the counting Bloom filter (CBF). The application of CBF enables us to structure candidate services into separate groups, resulting in an accelerated services discovery process. The contributions of this research work include a new approach to manage large number of services based on quantified service features, and a storage architecture design to support service discovery. Experimental results strongly support these claims.