Mobile ad hoc networks have gained a lot of attention lately as a means of providing continuous network connectivity to mobile computing devices regardless of physical location. Recently, a large amount of research has focused on the routing protocols needed in such an environment. In this two-part report, we investigate the effects that link breakage due to mobility has on TCP performance. Through simulation, we show that TCP throughput drops significantly when nodes move because of TCP's inability to recognize the difference between link failure and congestion. We also analyze specific examples, such as a situation where throughput is zero for a particular connection. We introduce a new metric, expected throughput, for the comparison of throughput in multi-hop networks, and then use this metric to show how the use of explicit link failure notification (ELFN) techniques can significantly improve TCP performance. In this paper (Part I of the report), we present the problem and an analysis of our simulation results. In Part II of this report, we present the simulation and results in detail.

Modern communication networks are able to respond to randomly uctuating demands and failures by adapting rates, by rerouting traffic and by reallocating resources. They are able to do this so well that, in many respects, large-scale networks appear as coherent, almost intelligent, organisms. The design and control of such networks present challenges of a mathematical, engineering and economic nature. This paper outlines how mathematical models are being used to address current issues concerning the stability and fairness of rate control algorithms for the Internet.

This note derives the stationary behavior of idealized TCP congestion avoidance. More specifically, it derives the stationary distribution of the congestion window size if loss of packets are independentevents with equal probability. The mathematical derivation uses a fluid flow, continuous time, approximation to the discrete time process #W n #, where W n is the congestion window after the n-th packet. We derive explicit results for the stationary distribution and all its moments. Congestion avoidance is the algorithm used by TCP to set its window size (and indirectly its data rate) under moderate to light segment (packet) losses. The congestion avoidance mechanism we model is idealized in the sense that loss of multiple packets does not lead to time-out phenomena. Such idealized behavior can be implemented using Selective Acknowledgements (SACKs). As such, our model predicts behavior of TCP with SACKs. It also is an approximate model in other situations. Among the results are that if eve...

Ad-hoc networks are completely wireless networks of mobile hosts, in which the topology rapidly changes due to the movement of mobile hosts. This frequent topology may lead to sudden packet losses and delays. Transport protocols like TCP have been built mainly for reliable, xed networks. Hence, when used in adhoc networks, TCP misinterprets this loss as congestion and invokes congestion control. This leads to unnecessary retransmissions and loss of throughput. To overcome this problem, a feedback scheme is proposed, so that the source can distinguish between route failure and network congestion. When a route is disrupted, the source is sent a Route Failure Noti cation(RFN) packet, allowing it to freeze its timers and stop sending packets. When the route is re-established, the source is informed through a Route Re-establishment Noti-cation (RRN) packet, upon which it resumes by unfreezing timers and continuing packet transmissions. The simulated performance of TCP on ad-hoc networks with and without feedback is compared and reported. It is observed that in the event of route failures, as the route re-establishment time increases, the use of feedback provides signi cant gains in throughput as well as savings in unnecessary packet transmissions. Several further enhancements and directions for future work are also sketched. 1

In today's networks, the evolution of wide-area services is constrained by standardization and compatibility concerns. The result is that the introduction of a new service occurs much more slowly than the emergence of new applications and technologies that benefit from it. To ameliorate this problem, an active network exploits mobile code and programmable infrastructure to provide rapid and specialized service introduction. A viable active network has the potential to change the way network protocols are designed and used, stimulating innovation and hastening the arrival of new functionality. There are, however, a number of challenges that must be overcome in the design of an active network. Chief among them are how to express new services as network programs, and how to execute these programs efficiently and securely.

We study in this paper a TCP-like linear-increase multiplicative -decrease flow control mechanism. We consider congestion signals that arrive in batches according to a Poisson process. We focus on the case when the transmission rate cannot exceed a certain maximum value. We write the Kolmogorov equations and we use Laplace Transforms to calculate the distribution of the transmission rate in the steady state as well as its moments. Our model is particularly useful to study the behavior of TCP, the congestion control mechanism in the Internet. By a simple transformation, the problem can be reformulated in terms of an equivalent M/G/1 queue, where the transmission rate in the original model corresponds to the workload in the `dual' queue. The service times in the queueing model are not i.i.d., and they depend on the workload in the system.

Abstract This paper describes the use of Congestion Pricing as a means of providing Congestion Control and Differentiated Quality of Service. The application of the proposed technique to the Internet Protocol has the advantage that it can be simply implemented using Explicit Congestion Notification. In particular: the network mechanism is independent of higher level protocols; the end systems can continue to exhibit current TCP behaviours; new multiprotocol flexibility is made available to end systems and users. Architectural issues are discussed, including important aspects of aggregation and charging. We describe our methodology for assessing the scheme via a distributed network simulator. Initial results are presented which compare and contrast various adaptive strategies that achieve a variable range of TCP-like behaviours. 1 Introduction The current control mechanisms of the Internet are not well suited to offering different Quality of Service (QoS) to different channels. TCP, the dominant transport control protocol, is appropriate for "besteffort " traffic, and aimed at data where lost information has to be retransmitted. However, other types of traffic, such as streamed video or audio, are sensitive to delay as well as packet loss. Such traffic is often carried over UDP, which has no generic feedback signals or error recovery. Certain commercial IP offerings make use of the TOS field to give a two-level service offering, though this is only possible when all routers are under the provider's control. The Diff-Serv proposals to provide a small number of QoS classes can be seen as an extension of this idea. Within the Diff-Serv community, Explicit Congestion Notification (ECN) has been proposed as means of triggering active queue management disciplines to implement congestion avoidance.

TCP congestion control [9] is designed for network stability, robustness and opportunistic use of network bu er and bandwidth resources on an end-to-end per-connection basis. Upon detecting packet loss, TCP infers congestion and trades o per-user goodput for network stability. Speci cally, TCP throughput can be approximated by a function which isinversely proportional to the round trip time, the timeout delays and the square root of loss probability [16]. While the use of packet loss as an indicator of congestion is a robust technique, packet loss itself has a profound e ect on performance { especially in terms of the variance in goodput seen by individual connections. This \fairness " problem also results in what is commonly known as the \World Wide Wait " experienced by a majority ofinteractive Internet applications such as WWW or ftp. Another auxiliary problem in TCP congestion control is the lack of control over bottleneck queueing delay due to the end-to-end nature of control. In this paper, we evaluate three proposed solutions for these problems- an improved drop scheme (RED), a bit-based explicit congestion noti cation scheme (ECN) and a scheme which explicitly and transparently controls TCP rate (Packeteer TCP rate control). Our studies indicate marked improvements in fairness as we move from RED through ECN to TCP rate control. All schemes control bottleneck queueing delay, but trade o other measures such asdrop rate, utilization and fairness, with TCP rate control exhibiting the best performance in terms of all metrics. In terms of deployment exibility, TCP rate control and RED allow widespread and immediate deployment because they are transparent to hosts (ECN is not because it requires TCP protocol modi cations). The minimal state requirements and protocol transparency of RED allows it a large deployment space. 1

An H # based robust controller is designed for a rate-feedback flow-control problem in single-bottleneck communication networks. The controller guarantees stability robustness to uncertain time-varying multiple time-delays in different channels. It also brings the queue length at the bottleneck node to the desired steady-state value asymptotically and satisfies a weighted fairness condition. Lower bounds for stability margins for uncertainty in the time-delays and for the rate of change of the time-delays are derived. A number of simulations are included to demonstrate the time-domain performance of the controller. Trade offs between robustness and time-domain performance are also discussed.

The problem of fair bandwidth sharing among adaptive (TCP) and non-adaptive (i.e. CBR-UDP) ows at an Internet gateway is considered. An algorithm that drops packet preventively, in an at-tempt to actively penalize the non-adaptive tra c that attempts to "steal " bu er space, and therefore bandwidth from the adaptive tra c ows, is presented. The algorithm maintains minimal ow state information and is therefore scalable. The performance of the algorithm is compared with other gateway algorithms and it is shown that, in the presence of non-adaptive tra c, it achieves a more balanced bandwidth allocation among the di erent ows. The behavior of a ow subjected to the given algorithm has also been analyzed in detail.