Traffic Adaptive Routing
Dynamic routing is known to induce oscillations and massive performance degradations. Therefore, routing in the Internet today is optimised only on coarse time scales, despite the bursty and highly volatile data traffic. In the main part of our work we describe a universally applicable distributed algorithm for dynamic routing/traffic engineering. It is based on a re-routing policy recently proposed in game theroiy. Extensive simulations with realistic bursty TCP traffic show that it converges quickly without oscillations, and reveal performance gains that outrival those of traditional methods. Further parts of the thesis describe a methodology for deriving global Web traffic demands, an analysis of Web search traffic, and a memory-efficient algorithm for obtaining ftraffic statistics.