Internet Traffic Management

Internet traffic volumes continue to grow at a great rate. For network operators it is important to understand and manage the traffic behaviour in order to meet service-level agreements with their costumers and to give end-users good communication performance.This thesis considers three aspects of Internet traffic management: web traffic modelling, bandwidth allocation to TCP flows, and traffic engineering. The areas all have in common the need to understand and handle Internet traffic behaviour. For web traffic modelling the goal of the work itself is to understand traffic behaviour and to be able to generate realistic traffic in simulations and lab experiments. For traffic engineering and bandwidth allocation to TCP flows the purpose is to develop methods to steer and control the traffic.The web is one of the most popular Internet applications. This thesis presents a simple model of web client traffic. Starting from a packet trace of web traffic, we derive empirical probability distributions describing session lengths, time between user requests for web pages, and the amount of data that is transferred due to a single user request. Using these probability distributions we implement a web-client traffic generator and show that the generated traffic has the same characteristics as the original web traffic, including the traffic variability.TCP is the predominant Internet transport protocol. The second aspect of traffic management in this thesis is dynamic allocation of bandwidth to TCP flows. TCP provides a reliable flow of data between two hosts and adapts its rate to the available capacity. Network technologies such as Dynamic synchronous Transfer Mode (DTM) provides channels with dynamically adjustable capacity. The issue is to adaptively allocate bandwidth to TCP flows, when both TCP and the bandwidth allocation scheme can react to changes in the network load. We use simulation to investigate the behaviour of a bandwidth allocation scheme, its effect on TCP flows and on a network that can vary its capacity.The objective of traffic engineering is to avoid congestion in the network and to make good use of available resources by controlling and optimising the routing. The challenge for traffic engineering in IP networks is to cope with the dynamics of Internet traffic demands. This thesis propose l-balanced routings that route the traffic on the shortest paths possible but make sure that no link is utilised to more than a given level l, if possible. L-balanced routing gives efficient routing of traffic and controlled spare capacity to handle unpredictable changes in traffic. We present an l-balanced routing algorithm based on multi-commodity flow optimisation. We also present a heuristic search method for finding l-balanced weight settings for the legacy routing protocols OSPF and IS-IS. We show that the search and the resulting weight settings work well in real network scenarios.