Development of a mobile web services discovery and composition model

Abstract

Mobile computing, a challenging paradigm aiming at using and providing services anywhere and anytime, places great challenges on dynamic service discovery and composition. These challenges are specially related to mobile context (terminal constraints, user mobility, network connection, etc.) In fact, the discovery model began with semantically rewriting and enriching services and user query with OWL-S ontology and its non-functional properties (context, QoWS and user preferences). Then a similarity calculation step has been performed between the enriched user query and the set of advertised web services to generate the most relevant ones. Furthermore, the resulting services will be the inputs of a mobile web service composition process. Mobile web service composition is based on a formal model aiming to compute dynamic context (Location, Bandwidth) uncertainty and web service sensitivity. These values contribute to select web services that will form the composition plan. The web service composition algorithm suggests to the end user, the most relevant plans sorted in ascending order according to their total sensitivity value. The implementation and comprehensive simulation experiments show the efficiency of the mobile web service discovery and composition models.