Simulation-based comparisons of Tahoe, Reno and SACK TCP

Fall, Kevin; Floyd, Sally

doi:10.1145/235160.235162

Cited by 885 publications

(546 citation statements)

References 10 publications

Supporting

Mentioning

527

Contrasting

Unclassified

Order By: Relevance

“…It just guarantees end-toend-service differentiation in the current Internet, i.e. two flows of different classes will get QoS accordingly, and flows of the same class will get almost the same service 13,14 . On the other side, as aggressiveness (i.e.…”

Section: Resultsmentioning

confidence: 99%

“…Besides, the differences between versions of TCP (Tahoe [12], Reno [13], New Reno [14], SACK [15]) only rely on the congestion control mechanisms, especially "fast retransmit" and "fast recovery" [3,16]. The evolution process of TCP, to improve its mean performance, dealt with increasing its aggressiveness according to available resources (of course trying to be as fair as possible) [13]. Recall that "aggressiveness" is defined in this paper as the ability for a protocol to consume resources.…”

Section: A Congestion Control Based Approachmentioning

confidence: 99%

“…-TCP Tahoe; -TCP New Reno; -TCP SACK (Selective ACKnowledgements) [13,15]; -TCP Vegas [18], a version of TCP, not deployed and up to our knowledge not used, that proposes a new congestion control mechanism. This congestion control mechanisms is based on the evaluation of the variation of the RTT on each packet.…”

Section: Charging Differentiated Congestion Control Classesmentioning

confidence: 99%

See 2 more Smart Citations

Coherent charging of differentiated services in the internet depending on congestion control aggressiveness

Owezarski

Larrieu

2003

Computer Communications

View full text Add to dashboard Cite

Abstract. This paper deals with both enforcing and charging end-to-end Quality of Service (QoS) and differentiated services in the Internet. Today, much work for optimizing QoS and defining charging mechanisms accordingly is related to low network layer (up to IP), and the current proposals under the spotlights are VPN, CDN, (over)-provisioning, and of course DiffServ. However, the service an user can get can be quite different from the one provided by the network layer (IP in the Internet). Transport protocols (TCP most of the time) induce oscillations (often seen as self-similarity) in the traffic, in particular because of their congestion control mechanisms. Based on this result, it is obvious that it is impossible to coherently charge such kinds of services at layer 3. QoS has also to be managed and enforced by transport protocols, that also have to strongly impact the way pricing is done. As a consequence, a new approach for services differentiation and charging (in addition to existing layer 3 approaches) at transport level is proposed. This approach relies on the aggressiveness of congestion control mechanisms because the more aggressive a protocol, the better the QoS it provides in case of a well provisioned network. And of course charging has to evolve accordingly. This idea is demonstrated by taking examples as UDP and several versions of TCP. In particular it is proved that service differentiation can be made that way in the Internet. This paper also gives a quantitative study of the QoS got by users depending on the amount of resources consumed, and it is observed that this function is linear. The charging of each service can then be quantified that way.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: A Congestion Control Based Approachmentioning

confidence: 99%

Section: Charging Differentiated Congestion Control Classesmentioning

confidence: 99%

See 1 more Smart Citation

Coherent charging of differentiated services in the internet depending on congestion control aggressiveness

Owezarski

Larrieu

2003

Computer Communications

View full text Add to dashboard Cite

show abstract

“…al. summarise the benefits of different improvements to the fast retransmit mechanism in the paper "Simulationbased comparisons of tahoe, reno, and sack tcp" [10]. TCP Tahoe implemented fast retansmit, thus being able to recover before a timeout was triggered.…”

Section: Related Workmentioning

confidence: 99%

Improving SCTP retransmission delays for time-dependent thin streams

Petlund

Beskow

Pedersen

et al. 2009

Multimed Tools Appl

View full text Add to dashboard Cite

A large number of network services rely on IP and reliable transport protocols. For applications that provide abundant data for transmission, loss is usually handled satisfactorily, even if the application is latency-sensitive (Wang et al. 2004). For data streams where small packets are sent intermittently, however, applications can occasionally experience extreme latencies (Griwodz and Halvorsen 2006). As it is not uncommon that such thin-stream applications are time-dependent, any unnecessarily induced delay can have severe consequences for the service provided. Massively Multiplayer Online Games (MMOGs) are a defining example of thin streams. Many MMOGs (like World of Warcraft and Age of Conan) use TCP for the benefits of reliability, in-order delivery and NAT/firewall traversal. It has been shown that TCP has several shortcomings with respect to the latency requirements of thin streams because of the way it handles retransmissions (Griwodz and Halvorsen 2006). As such, an alternative to TCP may be SCTP (Stewart et al. 2000), which was originally developed to meet the requirements of signaling transport. In this paper, we evaluate the Linux-kernel SCTP implementation in the context of thin streams. To address the identified latency challenges, we propose sender-side only enhancements that reduce the application-layer latency in a manner that is compatible with unmodified receivers. These enhancements can be switched on by applications and are used only when the system identifies the stream as thin. To evaluate the latency performance, we have performed several tests over various real networks and over an emulated network, varying parameters like RTT, packet loss and amount of competing cross traffic. When comparing our modifications with SCTP on Linux and FreeBSD and TCP New Reno, our results show great latency improvements and indicate the need for a separate handling of thin and thick streams.

show abstract

“…After a timeout of min{2 + L, 15} rounds, the flow restarts with a window equal to w/L. This penalty is justified by the experimental results of [7,19].…”

Section: The Window Gamementioning

confidence: 99%

Window-Games between TCP Flows

Efraimidis

Tsavlidis

2008

Algorithmic Game Theory

View full text Add to dashboard Cite

Abstract. We consider network congestion problems between TCP flows and define a new game, the Window-game, which models the problem of network congestion caused by the competing flows. Analytical and experimental results show the relevance of the Window-game to the real TCP game and provide interesting insight on Nash equilibria of the respective network games. Furthermore, we propose a new algorithmic queue mechanism, called Prince, which at congestion makes a scapegoat of the most greedy flow. Preliminary evidence shows that Prince achieves efficient Nash equilibria while requiring only limited computational resources.

show abstract

Simulation-based comparisons of Tahoe, Reno and SACK TCP

Cited by 885 publications

References 10 publications

Coherent charging of differentiated services in the internet depending on congestion control aggressiveness

Coherent charging of differentiated services in the internet depending on congestion control aggressiveness

Improving SCTP retransmission delays for time-dependent thin streams

Window-Games between TCP Flows

Contact Info

Product

Resources

About