Bandwidth assurance in a differentiated services network

The differentiated services (DS) architecture has been proposed for providing different levels of services and has recently received wide attention. A packet in a DS domain is classified into a class of service according to its contract profile and treated differently by its class. There are currently two types of service standardized, Expedited Forwarding (EF) and Assured Forwarding (AF). In this dissertation, we focus on AF service.; In AF service, customers make a contract in terms of bandwidth with their service providers. At ingress nodes, the service provider marks incoming packets IN or OUT based on the current sending rate and the contracted rate. At core routers, each packet observes different drop precedences according to its marking. With these simple marking and forwarding schemes, the service provider can provide different amounts of bandwidth to meet the customers' requests.; However, it has been observed that the drop precedences by themselves cannot achieve the desired rates because of the strong interaction of the transport protocol with packet drops in the network. In this dissertation, we propose and evaluate a number of techniques to better achieve the throughput guarantees in such networks. The proposed techniques consider modifying the transport protocol at the sender, the marking strategies at the marker and the dropping policies at the router. It is shown that these techniques improve the likelihood of achieving the desired throughput guarantees and also improve the service differentiation.; We also develop analytical models of TCP behavior in a DS network. The models quantitatively characterize TCP throughput as functions of the contract rate, the packet-drop rate and the round-trip time. The models analytically explain the previous observations such as unfair sharing of excess bandwidth among TCP flows with different contract rates.; We look at the problem of achieving specific QoS goals of individual flows by flexibly managing resources available to an aggregated source. It will be shown that an aggregated source can maintain a state of individual flows at the edge of the network and utilize this state effectively in adaptively marking packets of individual flows to meet their QoS goals.