Verifed Factorization Methods for SMP/G/1 Queueing Systems and their Interplay in an Integrated Problem-Solving Environment

Abstract

Semi-Markov processes (SMPs) are adequate models for correlated data traffic in service-integrated communication networks. In recent works, we have developed both modeling techniques and verified factorization methods for the analysis of queueing systems of GI/G/1 and SMP/G/1 types. These methods allow a quick and validated computation of workload distributions at the respective node. In this paper, we study enhancements and alternatives to these known methods. Our objective is to yield more accurate results for small SMP models, such as described previously in literature, as well as to be able to analyse larger SMP/G/1 models, for instance, resulting from modeling video traffic. These enhancements include a modified polynomial factorization technique as well as an application of the successive over-relaxation (SOR) technique to the verified Wiener-Hopf factorization. We give a brief overview of how we combine these methods in an integrated environment and the benefit for modeling and analysis of traffic in communication networks. Using this environment, we illustrate the particular techniques on two examples.

Extracted Key Phrases

3 Figures and Tables

Cite this paper

@article{Kempken2011VerifedFM, title={Verifed Factorization Methods for SMP/G/1 Queueing Systems and their Interplay in an Integrated Problem-Solving Environment}, author={Sebastian Kempken and Wolfram Luther}, journal={Reliable Computing}, year={2011}, volume={15}, pages={179-192} }