Selfish routing on the internet

Document Type

Syllabus

Publication Date

1-1-2004

Abstract

In recent years, we are more and more frequently dealing with large networks, like the Internet, that are not only computationally very complex, but they are also following highly nonstandard mechanisms and rules. The complex structure of such large networks is caused by their size and growth, their almost spontaneous emergence, their open architecture, and their lack of any authority that monitors and regulates network operations. Classical analyses of such networks have made assumptions either about some global regulation mechanisms or were designed with cooperative users in minds; in large networks nowadays both of these assumptions seem to be not very realistic. This initiated in recent years a growing interest in the analysis of large, noncooperative networks, where all the users decide by themselves on their own behavior. In such networks, each user is typically interested only in maximizing its own benefits at the lowest possible costs, and its choice of action will depend on those of other network’s users. The behavior of users in such an environment has been traditionally addressed in the framework of game theory that provides systematic tools to study and understand their socio-economic behavior, see, e.g., [1–8] and the references therein. A fundamental notion that arises in game theory is that of Nash equilibrium, which, informally, is the state where no user finds it beneficial to change it. Game theoretical models have been employed in networking in the context of traffic and bandwidth allocation, flow control, routing, etc. These studies typically investigate the properties of Nash equilibria and supply the insight into the nature of networking under decentralized and noncooperative control.

Identifier

84994848130 (Scopus)

ISBN

[9781584883975, 9780203489802]

Publication Title

Handbook of Scheduling Algorithms Models and Performance Analysis

First Page

42

Last Page

41

This document is currently not available here.

Share

COinS