DYNAMIC SEMI-MARKOVIAN WORKLOAD MODELING

Nima Sharifimehr, Samira Sadaoui

2008

Abstract

In this paper, we present a Markovian modeling approach which is based on a combination of existing Semi-Markov and Dynamic Markov models. The proposed approach is designed to be an efficient statistical modeling tool to capture both actions intervals patterns and sequential behavioral patterns. A formal definition of this model and detailed algorithms for its implementation are illustrated. We show the applicability of our approach to model workload of enterprise application servers. However, the given formal definition of our proposed approach prepares a firm ground for academic researchers to investigate many other possible applications. Finally, we prove the accuracy of our dynamic semi-Markovian approach for the most chaotic situations.

Paper Citation

in Harvard Style

Sharifimehr N. and Sadaoui S. (2008). DYNAMIC SEMI-MARKOVIAN WORKLOAD MODELING . In Proceedings of the Tenth International Conference on Enterprise Information Systems - Volume 1: ICEIS, ISBN 978-989-8111-36-4, pages 125-130. DOI: 10.5220/0001678801250130

in Bibtex Style

@conference{iceis08,
author={Nima Sharifimehr and Samira Sadaoui},
title={DYNAMIC SEMI-MARKOVIAN WORKLOAD MODELING},
booktitle={Proceedings of the Tenth International Conference on Enterprise Information Systems - Volume 1: ICEIS,},
year={2008},
pages={125-130},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0001678801250130},
isbn={978-989-8111-36-4},
}

in EndNote Style

TY - CONF
JO - Proceedings of the Tenth International Conference on Enterprise Information Systems - Volume 1: ICEIS,
TI - DYNAMIC SEMI-MARKOVIAN WORKLOAD MODELING
SN - 978-989-8111-36-4
AU - Sharifimehr N.
AU - Sadaoui S.
PY - 2008
SP - 125
EP - 130
DO - 10.5220/0001678801250130