QoS-Aware Service Selection for Abstract Workflows using Provenance Data and Fuzzy Constraint Satisfaction Modeling

Service-oriented applications are using services that most accurately meet their requirements; as a result Quality of Service (QoS)-based service selection mechanisms play an essential role in service-oriented architectures. In service-oriented environments such as the Grid, we usually define abstract workflows to enable the binding of services at runtime. Using service discovery techniques to discover services that functionally match tasks of the abstract workflow, enables to differentiate services in order to select the ones that better satisfy user QoS requirements. In this paper, we model the QoS-aware service selection for abstract workflows as a Constraint Satisfaction Problem (CSP) and demonstrate that soft CSPs like fuzzy CSP are an appropriate approach for these purposes. We provide evaluations of the model and discuss how effective the process of abstract workflow service selection can be when merging it with the degree of satisfaction. Furthermore, we exploit the hill climbing strategy using special heuristics to achieve a speedup in order to make the search process more effective and scalable.