Utilizing the Waiting-Time Criterion for Selecting Services in a Composition Scenario

Service composition is an effective practice to perform complex tasks through varied configurations of simple services. An issue that often arises in such a set-up is the selection of the best service from a group of functionally equivalent ones to cater to each of the various functionalities of the composite application. Previous effort in this direction incorporates utilizing the Quality of Service (QoS) attributes of the services to pick out the best one of the lot. This paper presents a technique to select the optimal service for composition using the average waiting time attribute of the services. The service selected is the one that has the smallest value of the average waiting time. Concepts from queueing theory have been borrowed and customized to estimate the waiting time values of the candidate services. Experiments have been performed wherein selection results using the proposed technique are compared with the selections that are made by simulating an actual scenario and computing the waiting time by observation.