Efficient Service Composition Using Zero-Suppressed Reduced Ordered Binary Decision Diagrams

Recent algorithms for automated service composition issue many complex queries to service directories. As service directories are shared resources, they may become performance bottlenecks. In order to increase scalability, we introduce a compact directory digest, which is distributed to clients and includes all information needed for automated service composition. Therefore, complex directory queries during service composition can be avoided. We encode a directory digest as a zero-suppressed reduced ordered binary decision diagram (ZDD). In several steps, we refine a simple service composition algorithm in order to leverage the ZDD representation. Introducing specialized ZDD operations, we achieve a service composition algorithm that scales very well with an increasing size of the directory digest