Transmission Control Protocol Flows Research Articles

An Explicit and Fair Window Adjustment Method to Enhance TCP Efficiency and Fairness Over Multihops Satellite Networks

Transmission control protocol (TCP) is the most widely used transport protocol in today's Internet. Despite the fact that several mechanisms have been presented in recent literature to improve TCP, there remain some vexing attributes that impair TCPs performance. This paper addresses the issue of the efficiency and fairness of TCP in multihops satellite constellations. It mainly focuses on the effect of the change in flows count on TCP behavior. In case of a handover occurrence, a TCP sender may be forced to be sharing a new set of satellites with other users resulting in a change of flows count. This paper argues that the TCP rate of each flow should be dynamically adjusted to the available bandwidth when the number of flows that are competing for a single link, changes over time. An explicit and fair scheme is developed. The scheme matches the aggregate window size of all active TCP flows to the network pipe. At the same time, it provides all the active connections with feedbacks proportional to their round-trip time values so that the system converges to optimal efficiency and fairness. Feedbacks are signaled to TCP sources through the receiver's advertised window field in the TCP header of acknowledgments. Senders should accordingly regulate their sending rates. The proposed scheme is referred to as explicit and fair window adjustment (XFWA). Extensive simulation results show that the XFWA scheme substantially improves the system fairness, reduces the number of packet drops, and makes better utilization of the bottleneck link.

Providing Differentiated Service to TCP Flows Over Bandwidth on Demand Geostationary Satellite Networks

The elasticity of transmission control protocol (TCP) traffic complicates attempts to provide performance guarantees to TCP flows. The existence of different types of networks and environments on the connections' paths only aggravates this problem. In this paper, simulation is the primary means for investigating the specific problem in the context of bandwidth on demand (BoD) geostationary satellite networks. Proposed transport-layer options and mechanisms for TCP performance enhancement, studied in the single connection case or without taking into account the media access control (MAC)-shared nature of the satellite link, are evaluated within a BoD-aware satellite simulation environment. Available capabilities at MAC layer, enabling the provision of differentiated service to TCP flows, are demonstrated and the conditions under which they perform efficiently are investigated. The BoD scheduling algorithm and the policy regarding spare capacity distribution are two MAC-layer mechanisms that appear to be complementary in this context; the former is effective at high levels of traffic load, whereas the latter drives the differentiation at low traffic load. When coupled with transport layer mechanisms they can form distinct bearer services over the satellite network that increase the differentiation robustness against the TCP bias against connections with long round-trip times. We also explore the use of analytical, fixed-point methods to predict the performance at transport level and link level. The applicability of the approach is mainly limited by the lack of analytical models accounting for prioritization mechanisms at the MAC layer and the nonuniform distribution of traffic load among satellite terminals.

A modification of TCP flow control for improving end-to-end TCP performance over networks with wireless links

End-to-end Transmission Control Protocol (TCP) performance is one of the more important issues in wireless Internet services. This paper proposes the improvement of end-to-end TCP performance via a TCP-aware link layer protocol called Adaptive TCP (A-TCP). The key idea behind the protocol is that an A-TCP agent, which is located in each base station, makes a mobile host look as if it has a wired link with the base station. This concept is referred to in this paper as the virtual host model. In order to implement this model, the A-TCP agent performs three functions: local retransmission, sender freezing and A-TCP flow control. A-TCP flow control is an original proposal and is also the principal factor for improving end-to-end TCP performance in a wireless Internet environment. In A-TCP flow control, the A-TCP agent marks the window field of each acknowledgment segment with a retransmission buffer size. Therefore, the TCP congestion controls, which happen in a TCP sender, are not caused by wireless link overflow. Performance evaluations were conducted via computer simulations and mathematical analyses. The results of the evaluations show not only that the A-TCP improves end-to-end TCP performance by at least 20%, which is higher compared to other TCP approaches, but also that the A-TCP can provide near-optimal performance in wireless bottleneck conditions. In addition, this paper proposes an A-TCP agent architecture, which consists of virtual host objects and several other functional blocks, and describes a new hand-off policy for the virtual host model, called transport-level soft hand-off.

Mobility support in hybrid wireless/IP networking

The ramifications of Transmission Control Protocol (TCP)/Internet Protocol (IP) networking with mobile hosts and wireless links in a hybrid wireless/IP environment are examined. Two issues that affect the performance of TCP/IP are mobility support in the network layer and coping with packet errors in TCP flow/congestion control in the transport layer. The operational features of agent-assisted, router-oriented and hybrid strategies for mobility support are identified and discussed. Packet errors due to transmission through the wireless channel and handover induced by user mobility can pose problem for the conventional TCP flow control in the transport layer. Strategies to cope with errors in packets that enter the TCP connection are described and discussed.

An effective feedback control mechanism for DiffServ architecture

As a scalable QoS (Quality of Service) architecture, DiffServ (Differentiated Service) mainly consists of two components: traffic conditioning at the edge of the DiffServ domain and simple packet forwarding inside the DiffServ domain. DiffServ has many advantages such as flexibility, scalability and simplicity. But when providing AF (Assured Forwarding) services, DiffServ has some problems such as unfairness among aggregated flows or among micro-flows belonging to an aggregated flow. In this paper, a feedback mechanism for AF aggregated flows is proposed to solve this problem. Simulation results show that this mechanism does improve the performance of DiffServ. First, it can improve the fairness among aggregated flows and make DiffServ more friendly toward TCP (Transmission Control Protocol) flows. Second, it can decrease the buffer requirements at the congested router and thus obtain lower delay and packet loss rate. Third, it also keeps almost the same link utility as in normal DiffServ. Finally, it is simple and easy to be implemented.

Analysis and design of controllers for AQM routers supporting TCP flows

In active queue management (AQM), core routers signal transmission control protocol (TCP) sources with the objective of managing queue utilization and delay. It is essentially a feedback control problem. Based on a recently developed dynamic model of TCP congestion-avoidance mode, this paper does three things: 1) it relates key network parameters such as the number of TCP sessions, link capacity and round-trip time to the underlying feedback control problem; 2) it analyzes the present de facto AQM standard: random early detection (RED) and determines that REDs queue-averaging is not beneficial; and 3) it recommends alternative AQM schemes which amount to classical proportional and proportional-integral control. We illustrate our results using ns simulations and demonstrate the practical impact of proportional-integral control on managing queue utilization and delay.

An optimization-oriented view of random early detection

To help improve the performance of congestion avoidance protocols like Transmission Control Protocol (TCP) and to limit the impact of non-adaptive UDP-based applications, the Internet Engineering Task Force has recommended the widespread deployment of active queue management schemes. We describe in this paper a new Random Early Detection (RED) algorithm for congestion control called Dynamic-RED (DRED) from a gradient optimization perspective. One of the goals of RED schemes is to stabilize the queue lengths in routers. However, the current version of RED does not succeed in this goal because the equilibrium queue length strongly depends on the number of active TCP connections. Using a simple optimization technique, DRED randomly discards packets with a load-dependent probability when a buffer in a router gets congested. DRED is also able to stabilize a router queue occupancy at a level independent of the number of active connections over a wide range of load levels. This is done without estimating the number of active TCP connections or flows and without collecting or analyzing state information on individual flows.