MatchCom: Stable Matching-Based Software Services Composition in Cloud Computing Environments

User preferences on throughput, latency, cost, service location, etc. indicate specific requirements when choosing a web service from the cloud marketplace. Service providers can also adopt preferences to prioritize a set of end-users based on their Service Level Agreement and service usage history. An effective matching between preferences from both parties enables fair service marketing in the cloud marketplace. The existing approaches are insufficient in capturing both parties’ preferences in the service composition process. To address this limitation, we propose MatchCom, a novel service composition approach driven by diverse preferences and formulate it as the stable marriage problem. Particularly, we present a novel fair preference ordering mechanism – in the context of a cloud marketplace, for enabling users to specify services provider ranking based on the capability they can provision, and for helping providers select the most suitable users to be served given users’ profile. MatchCom extends the Gale-Shapely Algorithm with a service composer algorithm for optimising the stable service composition. We evaluate MatchCom on a service-oriented system with 10 abstract services, each of which has 100 candidate web services. We establish through the experimental results that MatchCom outperforms other baseline approaches and can maximize end-user satisfaction in the composition process.