Reducing the complexity of the performance analysis of a multi- server facilities

Tulin Atmaca 1 Thomas Begin 2 Alexandre Brandwajn 3 Hind Castel-Taleb 1
2 DANTE - Dynamic Networks : Temporal and Structural Capture Approach
Inria Grenoble - Rhône-Alpes, LIP - Laboratoire de l'Informatique du Parallélisme, IXXI - Institut Rhône-Alpin des systèmes complexes
Abstract : Systems with multiple servers are common in many areas and their correct dimensioning is in general a difficult problem under realistic assumptions on the pattern of user arrivals and service time distribution. We present an approximate solution for the underlying Ph/Ph/c/N queueing model. Our approximation decomposes the solution of the Ph/Ph/c/N queue into solutions of simpler M/Ph/c/N and Ph/M/c/N queues. It is conceptually simple, easy to implement and produces generally accurate results for the mean number in the system, as well as the loss probability. A significant speed advantage compared to the numerical solution of the full Ph/Ph/c/N queue can be gained as the number of phases representing the arrival process and/or the number of servers increases.
Type de document :
Rapport
[Research Report] RR-8617, Institut Telecom, Telecom SudParis, Evry, France; Université Lyon 1 / LIP (UMR INRIA, ENS Lyon CNRS, UCBL), Lyon France; University of California Santa Cruz, Baskin School of Engineering, USA; INRIA. 2014
Liste complète des métadonnées

Littérature citée [14 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01076366
Contributeur : Thomas Begin <>
Soumis le : mardi 21 octobre 2014 - 20:15:01
Dernière modification le : jeudi 12 juillet 2018 - 01:09:29
Document(s) archivé(s) le : jeudi 22 janvier 2015 - 11:01:00

Fichier

RR-8617.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01076366, version 1

Citation

Tulin Atmaca, Thomas Begin, Alexandre Brandwajn, Hind Castel-Taleb. Reducing the complexity of the performance analysis of a multi- server facilities. [Research Report] RR-8617, Institut Telecom, Telecom SudParis, Evry, France; Université Lyon 1 / LIP (UMR INRIA, ENS Lyon CNRS, UCBL), Lyon France; University of California Santa Cruz, Baskin School of Engineering, USA; INRIA. 2014. 〈hal-01076366〉

Partager

Métriques

Consultations de la notice

285

Téléchargements de fichiers

157