Workflow-Based Service Selection under Multi-constraints

• Published in: 2015 IEEE International Conference on Services Computing pp. 332 - 339
• Main Authors: Chao Xia, Chi-Hung Chi, Wong, Raymond, Wombacher, Andreas, Pires, Luis F., van Sinderen, Marten, Chen Ding
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.06.2015
• Subjects: multi-constraints; scheduling; service computing; workflow
• DOI: 10.1109/SCC.2015.53

Despite the availability of services with similar functionality but from different providers in the cloud, using them in a workflow might subject to constraints such as service QoS and service bundling. Service bundling refers to the situation where the subscription of two services have to be done together, such requirement might be imposed by service providers and/or by the alliance group that the providers join in. In this paper, we focus on the service selection problem under the QoS constraints from the user and the bundling constraints associated with the chosen services. We first formulate the service selection problem as a multi-constrained selection problem. Then we propose a recursive heuristic search algorithm that takes the required QoS and bundling constraints into consideration for service selection. This algorithm has two unique functions: (i) utility function to measure the quality of the selection strategy under consideration, and (ii) acceptance function to limit the selection strategy only to those potential service candidates that have higher chance to satisfy the bundling constraints. Experiments show that our proposed solution can find better solutions than the existing ones without too much extra performance overhead.