On Graph Reduction for QoS Prediction of Very Large Web Service Compositions

Abstract : In this paper, we investigate the question of QoS prediction of Web Service Composition (WSC) implementing a business process. We focus on the graph reduction technique and the prediction of the Service Response Time. In the graph reduction technique, we assume that a Web Service Composition can be represented as a graph. The main thesis is that the QoS of such a graph graph can be obtained from a composition of the ones of its nodes. Multiple graph reduction algorithms have been proposed in the literature. Our contribution is twofold. We propose first a fast algorithm based on graph reduction for the prediction of the Service Response Time of a Web Service Composition. In comparison to those existing in the literature, this algorithm uses less memory space and has a better time complexity. The obtained improvements are in particular significant on very large Web Service Composition where the number of services is huge. Our second contribution is an analysis of the graph reduction technique for QoS prediction that takes into account the unfolding of services. In such cases, we show that the prediction of QoS can lead to a NP-complete problem. We also provide an integer programming model for predicting the Service Response Time in this case.
Document type :
Conference papers
Liste complète des métadonnées

https://hal.inria.fr/hal-00714138
Contributor : Gustavo Oliva <>
Submitted on : Wednesday, July 11, 2012 - 1:52:36 AM
Last modification on : Sunday, December 17, 2017 - 6:54:05 AM
Document(s) archivé(s) le : Friday, October 12, 2012 - 2:25:29 AM

File

SCC12-upload.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00714138, version 1

Collections

Citation

Alfredo Goldman, Yanik Ngoko. On Graph Reduction for QoS Prediction of Very Large Web Service Compositions. IEEE SCC - 9th International Conference on Service Computing - 2012, Jun 2012, Honolulu, United States. ⟨hal-00714138⟩

Share

Metrics

Record views

273

Files downloads

168