An efficient TCP flow control and fast recovery scheme for lossy networks

H. Y. Liao, Y. C. Chen, C. L. Lee

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The initial TCP Tahoe version uses the slow-start algorithm to deal with flow control and congestion avoidance. The later Reno version deploys both fast-retransmit and fast-recovery algorithms. Traditionally a segment loss is considered as owing to the network congestion. However, a packet loss may be caused by some other reason such as a transmission error in the wireless link. Due to this reason, we design a mechanism that subdivides the congestion control mechanism into two parts, the packet loss indication and the loss recovery. Regarding the former, we no longer treat the packet loss caused by the transmission error as an indication of network congestion. While for the latter, we proposed a modified scoreboard algorithm in TCP FACK to quickly recover the packet loss and prevent the retransmitted packet from being lost again.

Original languageEnglish
Title of host publicationNetworking - ICN 2001 - 1st International Conference on Networking, Proceedings
EditorsPascal Lorenz
PublisherSpringer Verlag
Pages82-96
Number of pages15
ISBN (Electronic)3540423036, 9783540423034
DOIs
StatePublished - 2001
Externally publishedYes
Event1st International Conference on Networking, ICN 2001 - Colmar, France
Duration: 09 07 200113 07 2001

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume2094
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference1st International Conference on Networking, ICN 2001
Country/TerritoryFrance
CityColmar
Period09/07/0113/07/01

Bibliographical note

Publisher Copyright:
© Springer-Verlag Berlin Heidelberg 2001.

Fingerprint

Dive into the research topics of 'An efficient TCP flow control and fast recovery scheme for lossy networks'. Together they form a unique fingerprint.

Cite this