Previous research on ECN and RED usually considered only a limited traffic domain, focusing on networks with a small number of homogeneous flows. The behavior of RED and ECN congestion control mechanisms in TCP network with many competing heterogeneous flows in the bottleneck link, hasn't been sufficiently explored. This thesis first investigates the behavior and performance of RED with ECN congestion control mechanisms with many heterogeneous TCP Reno flows using the network simulation tool, ns-2. By comparing the simulated performance of RED and ECN routers, this study finds that ECN does provide better goodput and fairness than RED for heterogeneous flows. However, when the demand is held constant, the number of flows generating the demand has a negative effect on performance. Meanwhile, the simulations with many flows demonstrate that the bottleneck router's marking probability must be aggressively increased to provide good ECN performance. Based on these simulation results, an Adaptive ECN algorithm (AECN) was studied to further improve the goodput and fairness of ECN. AECN divides all flows competing for a bottleneck into three flow groups, and deploys a different max for each flow group. Meanwhile, AECN also adjusts min for the robust flow group and max to get higher performance when the number of flows grows large. Furthermore, AECN uses mark-front strategy, instead of mark-tail strategy in standard ECN. A series of AECN simulations were run in ns-2. The simulations show clearly that AECN treats each flow fairer than ECN with the two fairness measurements: Jain's fairness index and visual max-min fairness. AECN has fewer packet drops and alleviates the lockout phenomenon and yields higher goodput than ECN.

The growth of high speed Internet connections has fueled an increase in the demand for high quality streaming video. In order to satisfy timing constraints, streaming video typically uses UDP as the default network transport protocol. Unfortunately, UDP does not have any end-to-end congestion control mechanisms, and so in the absence of higher layer congestion control can lead to unfairness and possibly congestion collapse. While there has been research done in video measurement and characterization using custom tools, to the best of our knowledge, there have been no measurement studies where the researchers had control over a commercial streaming media server and client, and control of the network conditions and content. A goal of this research is to characterize the bitrate response of Windows Streaming Media in response to network-level metrics such as capacity, loss rate, and round-trip time. We build a streaming media test bed that allows us to systematically vary network and content encoding characteristics. We analyze responsiveness by comparing streaming media flows to TCP-friendly flows under various streaming configurations and network conditions. We find Windows Streaming Media has a prominent buffering phase in which it sends data at a bitrate significantly higher than the steady-state rate. Overall, Windows Streaming Media is responsive to available capacity, but is often unfair to TCP. Knowledge of streaming media's response to congestion encountered in the network is important in building networks that better accommodate their turbulence. The additional characteristics we measure can be combined to guide emulation or simulation configurations and network traffic generators for use in further research.

An increasing number of network applications use reliable transport protocols. Applications with constant data transmission recover from loss without major performance disruption, however, applications that send data sporadically, in small packets, also called thin-streams, experience frequently high latencies due to 'Bufferbloat', that reduce the application performance. Active Queue Management mechanisms were proposed to dynamically manage the queues in routers by dropping packets early and reduce these, hence reducing latency. While their deployment to the internet remains an open issue, the proper investigation into how their functioning mechanism impacts latency is the main focus of this work and research questions have been devised to investigate the AQM impact on latency. A range of AQM mechanisms has been evaluated by the research, exploring performance of the methods for latency sensitive network applications. This has explored new single queue AQM mechanisms such as Controlled Delay (CODEL) and Proportional Integral Enhanced (PIE) and Adaptive RED (ARED). The evaluation has shown great improvements in queuing latency when AQM are used over a range of network scenarios. Scheduling AQM algorithms such as FlowQueue CODEL (FQ-CODEL) isolates traffic and minimises the impact of Bufferbloat on flows. The core components of FQ-CODEL, still widely misunderstood at the time of its inception, have been explained in depth by this study and their contribution to reducing latency have been evaluated. The results show significant reductions in queuing latency for thin streams using FQ-CODEL. When TCP is used for thin streams, high application latencies can arise when there are retransmissions, for example after dropping packets by an AQM mechanism. This delay is a result of TCP's loss-based congestion control mechanism that controls sender transmission rate following packet loss. ECN, a marking sender-side improvement to TCP reduces applicationlayer latency without disrupting the overall network performance. The thesis evaluated the benefit of using ECN using a wide range of experiments. The findings show that FQ-CODEL with ECN provides a substantial reduction of application latency compared to a drop-based AQM. Moreover, this study recommends the combination of FQ-CODEL with other mechanisms, to reduce application latency. Mechanisms such as ABE, have been shown to increase aggregate throughput and reduce application latency for thin-stream applications.

The Internet Engineering Task Force (IETF) has recently released network mobility standards that allow deployment of TCP/IP networks onboard a vehicle and maintain permanent network connectivity to the Internet via a vehicular mobile router. This recent development opens up new opportunities for providing efficient mobile computing for users on the move, especially for commuters traveling on public transports. Moreover, central and coordinated management of mobility in a single router, rather than by each user device individually, has numerous advantages. In this architecture, however, it becomes challenging to guarantee network performance due to the mobility of the network and inherently vulnerable nature of wireless links. In this thesis, a detailed performance study of onboard networks is conducted. It has been shown that disruptions in the mobile router connectivity can significantly degrade network throughput. Moreover, factors such as the limited wireless bandwidth of the access link, variations in the bandwidth due to technology switching, and the communication diversity of onboard users all contribute to the problem of unfair sharing of wireless bandwidth. By leveraging the fact that all onboard communications go through the mobile router, performance enhancing solutions are proposed that can be deployed in the mobile router to transparently address the throughput and fairness problems. In this architecture, when the route is known in advance and repetitive (e.g. for public transport or a regularly commuting private vehicle), a certain degree of prediction of impending link disruptions is possible. An anticipatory state freezing mechanism is proposed that relies on the prediction of link disruptions to freeze and unfreeze the state machine of TCP, the widely used transport protocol in the Internet. Simulation study shows that TCP throughput has a non-linear relationship with the prediction accuracy. As prediction accuracy increases, throughput problem diminishes quickly. An adaptive mobile router based fairness control mechanism is proposed to address the unfair sharing of wireless bandwidth in highly dynamic scenarios. The fairness is controlled by dynamically estimating the round-trip-times of all onboard TCP connections and transparently adjusting the protocol control parameters at the router. The thesis also discusses implementation issues for the proposed solutions.

Emerging wireless networks are characterized by increased heterogeneity in wireless access technologies as well as increased peer-to-peer communication among wireless hosts. The heterogeneity among wireless access interfaces mainly exists because of the fact that different wireless technologies deliver different performance trade-offs. Further, more and more infrastructure-less wireless networks such as ad-hoc networks are emerging to address several application scenarios including military and disaster recovery. These infrastructure-less wireless networks are characterized by the peer-to-peer communication model. In this thesis, we propose transport protocols that tackle the challenges that arise due to the above-mentioned properties of state-of-the-art wireless data networks. The main contributions of this work are as follows: 1. We determine the ideal nature and granularity of transport adaptation for efficient operation in heterogeneous wireless data networks by performing comprehensive experimental analysis. We then design and implement a runtime adaptive transport framework, *TP, which accommodates the capabilities of the ideal transport adaptation solution. 2. We prove that conversational transport protocols are not efficient under peer-to-peer wireless data networks. We then design and implement NCTP which is a non-conversational transport protocol.

Thesis (M.S.)--Mississippi State University. Department of Electrical and Computer Engineering. / Title from title screen. Includes bibliographical references.

Thesis (M.S.)--Worcester Polytechnic Institute. / Keywords: PI; PISA; PIMC; cwnd; TCP. Includes bibliographical references (p. 49-50).

Thesis (Ph. D.)--Mississippi State University. Department of Computer Science and Engineering. / Title from title screen. Includes bibliographical references.

Thesis (M.S.) -- Mississippi State University. Department of Electrical and Computer Engineering. / Title from title screen. Includes bibliographical references.

The Transmission Control Protocol (TCP) is the most widely used transport protocol in the Internet. TCP is a reliable transport protocol that is tuned to perform well in wired networks where packet losses are mainly due to congestion. Wireless channels are characterized by losses due to transmission errors and handoffs. TCP interprets these losses as congestion and invokes congestion control mechanisms resulting in degradation of performance. TCP is usually layered over the Internet protocol (lP) at the network layer. JP is not reliable and does not provide for any Quality of Service (QoS). The Internet Engineering Task Force (IETF) has provided two techniques for providing QoS in the Internet. These include Integrated Services (lntServ) and Differentiated Services (DiffServ). IntServ provides flow based quality of service and thus it is not scalable on connections with large flows. DiffServ has grown in popularity since it is scalable. A packet in a DiffServ domain is classified into a class of service according to its contract profile and treated differently by its class. To provide end-to-end QoS there is a strong interaction between the transport protocol and the network protocol. In this dissertation we consider the performance of the TCP over a wireless channel. We study whether the current TCP protocols can deliver the desired quality of service faced with the challenges they have on wireless channel. The dissertation discusses the methods of providing for QoS in the Internet. We derive an analytical model for TCP protocol. It is extended to cater for the wireless channel and then further differentiated services. The model is shown to be accurate when compared to simulation. We then conclude by deducing to what degree you can provide the desired QoS with TCP on a wireless channel. / Thesis (M.Sc.Eng.)-University of Natal, Durban, 2001.