5th International ICST Conference on Performance Evaluation Methodologies and Tools

Research Article

Bounding techniques for transient analysis of G-networks with catastrophes

Download563 downloads
  • @INPROCEEDINGS{10.4108/icst.valuetools.2011.245719,
        author={Hind Castel and Idriss Ismael-Aouled and Nihal  Pekergin},
        title={Bounding techniques for transient analysis of G-networks with catastrophes},
        proceedings={5th International ICST Conference on Performance Evaluation Methodologies and Tools},
        publisher={ICST},
        proceedings_a={VALUETOOLS},
        year={2012},
        month={6},
        keywords={G-Networks transient analysis stochastic comparisons},
        doi={10.4108/icst.valuetools.2011.245719}
    }
    
  • Hind Castel
    Idriss Ismael-Aouled
    Nihal Pekergin
    Year: 2012
    Bounding techniques for transient analysis of G-networks with catastrophes
    VALUETOOLS
    ICST
    DOI: 10.4108/icst.valuetools.2011.245719
Hind Castel1,*, Idriss Ismael-Aouled2, Nihal Pekergin2
  • 1: Telecom SudParis
  • 2: LACL, Univ. PAris-Est Creteil
*Contact email: hind.castel@it-sudparis.eu

Abstract

We apply stochastic comparisons in order to bound the transient behavior of G-networks with catastrophes. These networks belong to Gelenbe's networks, with both positive and negative customers (or signals). We consider catastrophes where the signal deletes all customers in a queue. G-networks have a known product form steady-state distribution, but it is still impossible to obtain the transient distributions by a closed form. In the present paper, we propose to define smaller queueing systems providing bounds for subnetworks of the G-network with catastrophes. We apply stochastic comparisons by mapping functions to build bounding models. We derive transient performance measure bounds for applications as malware software infections. For instance, we obtain bounds for the first time of infection, or the number of times a station has been infected in a time interval. We study the tradeoff between the size of the subnetwork and the quality of the bounds with respect to parameters.