Splitting TCP Connections Adaptively Inside Networks

Kivs, 2005

Wide area mobile networks facilitate TCP/IP with radio link protocols (RLP) in order to achieve acceptable throughput. Varying channel properties, roaming etc. which interfere with the TCP retransmission and congestion control mechanisms can be alleviated by split connection techniques and performance enhancing proxies (PEP). However, current split connection techniques and PEP do not sufficiently maintain end-to-end congestion control and rate control. Flow control and window clamping techniques, which are often used for this purpose, do not solve but work around the problem and undermine TCP's sliding window mechanism. In this paper we propose the concept of Path Tail Emulation to overcome this weakness. A TCP flow is split at the gateway from the Internet into the mobile network. The mobile network is then hidden behind an emulated loss free link, the bandwidth and capacity of which correspond to the mobile network's properties. Using Path Tail Emulation, congestion control and acknowledgement pacing is done exactly the same way as in pure wirebound networks, thus TCP can fully exploit the available network capacity without suffering from any adverse interaction with lower layers. The behaviour of a mobile network behind a PEP using Path Tail Emulation is reduced to a network which is compliant to the network model used for TCP in wirebound networks and therefore can be handled with well known and well proven techniques.

As the Internet is expected to better support many applications such as multimedia with limit bandwidth, new mechanisms are needed to control the congestion in the network. Congestion control plays the key role to ensure stability of the Internet along with fair and efficient allocation of the bandwidth. So, congestion control is currently a large area of research and concern in the network community. Many congestion control mechanisms are developed and refined by researcher aiming to overcome congestion. During the last decade, several congestion control mechanisms have been proposed to improve TCP congestion control.

ACM SIGMETRICS Performance Evaluation Review, 2000

Our research focuses on end-to-end congestion avoidance algorithms that use round trip time (RTT) fluctuations as an indicator of the level of network congestion. The algorithms are referred to as delay-based congestion avoidance or DCA. Due to the economics associated with deploying change within an existing network, we are interested in an incrementally deployable enhancement to the TCP/Reno protocol. For instance, TCP/Vegas, a DCA algorithm, has been proposed as an incremental enhancement. Requiring relatively minor modifications to a TCP sender, TCP/Vegas has been shown to increase end-to-end TCP throughput primarily by avoiding packet loss. We study DCA in today's best effort Internet where IP switches are subject to thousands of TCP flows resulting in congestion with time scales that span orders of magnitude. Our results suggest that RTT-based congestion avoidance may not be reliably incrementally deployed in this environment. Through extensive measurement and simulation, ...

IEEE Communications Surveys & Tutorials, 2003

2013 IEEE Global Communications Conference (GLOBECOM), 2013

The bandwidth utilization in traditional TCP protocols (e.g., TCP New Reno) suffers over high-latency and highbandwidth links due to the inherent characteristics of TCP congestion control. Conventional methods of improving throughput cannot be applied per se for streaming applications. The challenge is exacerbated by "big data" applications such as with the Long Wavelength Array data that is generated at a rate of up to 4 terabytes per hour. To improve bandwidth utilization, we introduce layer-4 relay(s) that enable the pipelining of TCP connections. That is, a traditional end-to-end connection is split into independent streams, each with shorter latencies, that are then concatenated (or cascaded) together to form an equivalent end-to-end TCP connection. This addresses the root cause by decreasing the latency over which the congestion-control protocol operates. To understand when relays are beneficial, we present an analytical model, empirical data and its analyses, to validate our argument and to characterize the impact of latency and available bandwidth on throughput. We also provide insight into how relays may be setup to achieve better bandwidth utilization.

Originally TCP was designed for early, low bandwidth, short distance networks, so Standard TCP did not utilize the maximum bandwidth in today's high bandwidth network environments. Therefore a lot of TCP congestion control mechanisms also known as TCP variants have been developed for today's long distance high bandwidth networks. In this paper the experimental results evaluating the performance of TCP Reno, HighSpeed TCP, BIC TCP, TCP CUBIC and Compound TCP in short and long distance high bandwidth networks are presented. Results show that TCP CUBIC shows the highest performance in goodput whereas TCP Compound shows the highest performance in protocol fairness and TCP friendliness as compared to the other stat of the art congestion control mechanisms.

International Journal of Computer and Communication Engineering, 2013

A reliable end to end communication is a buzzword that is promised by the transport layer protocol TCP. TCP, a Reliable transport protocols are tuned to perform well in different networks but, packet losses occur mostly because of congestion. TCP contains several mechanisms (such as slow start, congestion avoidance, fast retransmit and fast recovery) for ensuring reliability. However, it has reached its limitation in some challenging network environments like-High speed communication, Communication over different media. Thus, it requires further analysis and development of congestion control algorithms. In this paper, we have explored the reliability and robustness of TCP variants (Tahoe, Reno, New-Reno, SACK, FACK and TCP VEGAS, HSTCP, CUBIC TCP) based on different parameters such as throughput, end-to-delay, jitter and packet drop ratio over wired and wireless networks. We have also compared and discussed different congestion control and avoidance mechanisms of TCP variants to show how they affect the throughput and efficiency of different network environments.

In this paper, we propose an endpoint congestion management scheme, called COordinated COngestion cONtrol (COCOON). The basic idea is to identify and group connections that may traverse the same backbone link, to enable them to share congestion information, and to coordinate among them all the congestion avoidance/control activities. The size of a COCOON group can be dynamically adjusted so as to magnify the benefits of end-host congestion management. COCOON also allows a new connection to commence with a congestion window that is large enough to catch up with other connections while not inducing congestion. Finally, COCOON takes into account nonresponsive UDP connections and "bundles" them into a virtual connection that is subject to TCP-like congestion control. significantly reduces the packet loss ratio of concurrent connections, while sustaining the throughput comparable to the best scheme. This, coupled with the fact that COCOON requires only minor modification at server hosts, suggests that COCOON can be readily deployed over the Internet.

Congestion control remains an important topic for today’s Internet protocols. Congestion is generally bad for users, applications and networks. Several mechanisms were proposed by researchers to improve congestion control. These mechanisms include TCP Tahoe, Reno, Vegas, SACK, and NewReno. In this paper, we evaluate the current congestion control protocols considering throughput, losses, delay, and fairness that provided by each variant. This study is done using the well known network simulator NS-2 and a realistic topology generator called GT-ITM.