TCP Vegas: end to end congestion avoidance on a global Internet

Brakmo, Lawrence S.; Peterson, Larry L.

doi:10.1109/49.464716

Cited by 1,360 publications

(838 citation statements)

References 14 publications

Supporting

Mentioning

804

Contrasting

Unclassified

Order By: Relevance

“…Several bandwidth estimation schemes have been proposed in the literature for the TCP congestion control [4,5,6]. Here we focus our analysis mainly over TCP Westwood, which was recently proposed in [3].…”

Section: Estimation Algorithmsmentioning

confidence: 99%

“…Clustering [7,8], which is experimented by TCP segments belonging to the same connection; ACK Compression [7,8], which alters the spacing between received ACKs due to congested routers in the retourning path, thus causing possible estimation problems; TCP coarse-grained clocks [6,9], i.e. the low precision with which many real TCP implementations take measures of the RTT.…”

Section: Estimation Problemsmentioning

confidence: 99%

“…More detailed descriptions of that implemented by TCP Vegas is in [6], and of that adopted by TCP Westwood is in [3].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Enhanced Bandwidth Estimation Algorithms in the TCP Congestion Control Scheme

Capone

Fratta

Martignon

2003

IFIP Advances in Information and Communication Technology

View full text Add to dashboard Cite

The use of enhanced bandwidth estimation procedures within the congestion control scheme of TCP has been recently proposed as a way to improve the performance over links affected by random losses. In this paper we propose TIBET (Time Intervals based Bandwidth Estimation Technique), a new bandwidth estimation scheme that can be implemented within the TCP congestion control procedure by modifying only the sender-side of a connection. TIBET allows TCP sources to achieve higher performance over wireless links, while guaranteeing a good fairness level with TCP Reno over wired networks. We analyze and compare the performance of the new algorithm with that of previously proposed schemes.

show abstract

Section: Estimation Algorithmsmentioning

confidence: 99%

Section: Estimation Problemsmentioning

confidence: 99%

See 1 more Smart Citation

Enhanced Bandwidth Estimation Algorithms in the TCP Congestion Control Scheme

Capone

Fratta

Martignon

2003

IFIP Advances in Information and Communication Technology

View full text Add to dashboard Cite

show abstract

“…These observations, together with the desire to keep the congestion control algorithm in end-hosts only, without requiring the cooperation of the network, lead to propose new versions of TCP, like Vegas [2] or Westwood [11], that try to estimate the available bandwidth (or fair bandwidth share) within the network. In both schemes the major problem arises from the difficulty of correctly estimating the available bandwidth.…”

Section: Tcp Protocol Modificationmentioning

confidence: 99%

Sender- Side TCP Modifications: An Analytical Study

Cigno

Procissi

Gerla

2002

NETWORKING 2002: Networking Technologies, Services, and Protocols; Performance of Computer and Communication Networks; Mobile A

View full text Add to dashboard Cite

Abstract. This paper considers a number of modifications that can be applied to the congestion control algorithm of a TCP sender without requiring the cooperation either of the network or of the receiver, analyzing their impact on the performance of the protocol. We use a theoretical approach based on the use of queueing networks for the description of the protocol dynamics and a fixed point approximation to derive the working point of the IP network. Our results show that in presence of short lived connections the impact of the transient behavior of TCP on the network performance is dominant, and major performance improvements can be obtained only if the transient behavior is improved.

show abstract

“…These can negate the gain in network performance. Other approaches such as Tri-S [6] and TCP-Vegas [7] attempt to estimate the bandwidth-delay product for each TCP connection and adjust the window size based on this estimate. However, these schemes introduce complexity in the end-system and require extensive modifications to current TCP implementations.…”

Section: Related Workmentioning

confidence: 99%

Fair Intelligent Congestion Control Resource Discovery Protocol on TCP Based Network

Hoang

et al. 2003

Converged Networking

View full text Add to dashboard Cite

Abstract:Today's Internet only provides best-effort service for all traffics. The network is not able to guarantee the quality of service required by an application that demands more stringent response in terms of delay, jitters, bandwidth and etc. It is weil accepted that, the deployment of QoS-aware technologies is a key factor for the continued success of the Internet. In this paper, we propose a Fair Iiltelligent Congestion Control Resource Discovery (FICCRD) protocol on TCP based network whereby a mechanism is employed at core routers to determine available network resources and convey this information to edge routers. At the edge routers, an intelligent control algorithm is employed to assist the TCP to maximize its traffic over the underlying network. The key ideas are to integrale available network resources in estimating connections' fair share; to create feedback controlloops between edge routers; to employ a special Resource Discovery (RD) packet to collect and convey en route router state information; and to employ intelligent algorithms to match a TCP connection's sending rate to the rate at which the underlying network can support. We demonstrate that FICCRD protocol is effective, fair, flexible and can be easily extended for QoS control of the future Internet.Key words: TCP, Congestion Control, Explicit Window Adaptation, Acknowledgment Bucket Control, Feedback Control LoopThe original version of this chapter was revised: The copyright line was incorrect. This has been corrected. The Erratum to this chapter is available at DOI: INTRODUCTIONToday's Internet only provides best-effort service for all traffic. Traffic is processed as quickly as possible but there is no guarantee as to timeliness or actual delivery. The network makes no attempt to differentiale its service response between the traffic streams generated by concurrent users of the network. This means that the network is not able to guarantee the Ievel of service required by an application that demands more stringent response in terms of delay, jitters, bandwidth, etc. A simple application such as videoconference over the Internet will quickly demonstrate the problems with real-time applications. A person taking part in a videoconference will immediately realize that the quality of the images is not as good as expected, the flow of images is not quite smooth and the synchronization between voice and images is far from perfect.The evolution of the Internet is at a tuming point. It is well accepted that, in its evolution in the 21st century, the deployment of Quality of Service (QoS) aware technologies is a key factor for the continued success of the Internet. Some advances in this direction have already been provided with the introduction of the Internet augmented architectures with QoS supports: the Internet Protocol Version 6 (IPV6), the Despile these advances, many issues still remain to be solved. Among them, issues concerning admission control and resource allocation policies in the network, adequate user signalling protocols, networ...

show abstract

TCP Vegas: end to end congestion avoidance on a global Internet

Cited by 1,360 publications

References 14 publications

Enhanced Bandwidth Estimation Algorithms in the TCP Congestion Control Scheme

Enhanced Bandwidth Estimation Algorithms in the TCP Congestion Control Scheme

Sender- Side TCP Modifications: An Analytical Study

Fair Intelligent Congestion Control Resource Discovery Protocol on TCP Based Network

Contact Info

Product

Resources

About