Title: Experiences of sampling-based approaches for estimating QoS parameters in the Web Service composition problem

Authors: Eduardo Blanco; Yudith Cardinale; María-Esther Vidal

Addresses: Universidad Simón Bolívar, Valle de Sartenejas, Apdo 89000, Caracas, 1080-A, Venezuela ' Universidad Simón Bolívar, Valle de Sartenejas, Apdo 89000, Caracas, 1080-A, Venezuela ' Universidad Simón Bolívar, Valle de Sartenejas, Apdo 89000, Caracas, 1080-A, Venezuela

Abstract: QoS parameters can be used to characterise services or their compositions according to non-functional criteria. We propose sampling-based techniques to accurately estimate QoS values that will be used in a hybrid composer named PT-SAM, to identify the compositions that best meet a user request. PT-SAM adapts a Petri-Net unfolding and uses a utility function defined on the QoS estimates, to guide the composer. We report on the quality of the estimates and the performance of the composer. Our experiments show that solutions generated by PT-SAM are close to the optimal solutions while the composer can scale to large search spaces.

Keywords: web service composition; query optimisation; semantic matching; estimation techniques; QoS estimation; quality of service; web services; Petri nets.

DOI: 10.1504/IJWGS.2012.046729

International Journal of Web and Grid Services, 2012 Vol.8 No.1, pp.1 - 30

Published online: 31 Dec 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article