Towards a Comprehensive Model for the Study of Computer Networks Performance using Information Theory

Abstract

Improving computer networks performance is a tremendous need in our life for meeting ever-changing quality of service (QoS) requirements of interesting applications. We are searching continuously on how to increase the performance of computer networks. One of the most important factors that aid in determining network performance is the available and accessible network information. In this paper, we deal with this problem in an information theoretic framework. Our approach is different than most recent papers in this field, where we use predefined measures that normally work on any computer network and we apply it to a general network model. We examine these measures on a specific model that is more comprehensive and more applicable than previous models discussed in other papers. Also, one of the important reasons in our extended study is to investigate the shortcomings in adopting these measures and how they appear in the results. We discuss briefly how to enhance these measures to be compatible with real networks and to give more precise and efficient results. Our analytical study of the developed model shows that obtained results can be used to guide future network designs since our model can be easily manipulated to get many real network models.