Modeling of Network Packet Switches Using Matrix Analysis

Authors

Oladele Theophilus Sule, Department of Electrical and Computer Engineering, New Jersey Institute of Technology, Newark, NJ 07102, USAFollow
Roberto Rojas-Cessa, Department of Electrical and Computer Engineering, New Jersey Institute of Technology, Newark, NJ 07102, USAFollow

Author Country (or Countries)

USA

Abstract

We introduce the use of matrix analysis as a method for describing the operation of a packet switch in this paper. Matrix analysis may be used to statistically describe how incoming traffic is switched from inputs to outputs according to the architecture of the switch and the selection of packets for forwarding. This method finds a direct application on throughput analysis of a packet switch, defined by the switch architecture and configuration scheme. We show the applicability of this method on different known switch architectures, as examples, and show that the results are consistent with their reported performance in the literature. Specifically, the method is applied to single- and multi-stage switches using load-balancing or matching and arbitration schemes in their configuration process. Moreover, the method is used to obtain the throughput of a switch and its corresponding configuration scheme for a given doubly stochastic input matrix.

Digital Object Identifier (DOI)

http://dx.doi.org/10.18576/amis/140103

Recommended Citation

Theophilus Sule, Oladele and Rojas-Cessa, Roberto (2020) "Modeling of Network Packet Switches Using Matrix Analysis," Applied Mathematics & Information Sciences: Vol. 14: Iss. 1, Article 3.
DOI: http://dx.doi.org/10.18576/amis/140103
Available at: https://digitalcommons.aaru.edu.jo/amis/vol14/iss1/3