When is it Convenient to Predict the Web Services Completion Time?

When multiple distributed instances of the same Web service exist, a great challenge is represented by allowing the service requesters to select the most convenient instances at execution time. In the Web service scenario, the requesters represent the critical decision making point about what service instance to use. For this reason they need information to estimate the behavior of different services to perform an aware selection, instead of a random one. This work formally defines the select logical operation as a new step of the bind logical operation. Then it contributes to the comprehension of the conditions under which is convenient, for a service requestor, to spend a little time using prediction mechanisms for collecting service convenience values. Moreover we propose some simulation results obtained modelling a realistic Web service scenario.