A heuristic approach to fair routing path selection

Abstract

Knowledge about efficient usage of network resources is an important prerequisite for user collaboration in a static networked environment. Most of the network infrastructure internals cannot be directly sensed by users, except traffic allowances and response times. Here we will provide a technical approach for demonstrating to the user, how much changing traffic flows can guide to better employment of the shared network resources for all users together. The approach is based on order theory and fairness relations. Using the maxmin fair dominance relation, from its maximum set of maxmin fair states of random routings, the element with the largest throughput will be chosen, to yield a fair routing. The reported experiments will demonstrate that, for example, for a group of about 10 users, 500 samples are sufficient to report a suitable fair traffic allocation to the users. © 2011 Springer-Verlag.