DSpace
 

EPrints@IIT Delhi >
Faculty Research Publicatons  >
Computer Science and Engineering >

Please use this identifier to cite or link to this item: http://eprint.iitd.ac.in/handle/2074/2236

Full metadata record

DC FieldValueLanguage
dc.contributor.authorChrungoo, Abhay-
dc.contributor.authorGupta, Vishu-
dc.contributor.authorSaran, Huzur-
dc.contributor.authorShorey, Rajeev-
dc.date.accessioned2006-10-12T06:49:47Z-
dc.date.available2006-10-12T06:49:47Z-
dc.date.issued2001-
dc.identifier.citationGlobal Telecommunications Conference, GLOBECOM IEEE, 3, 1713 - 1717p.en
dc.identifier.urihttp://eprint.iitd.ac.in/dspace/handle/2074/2236-
dc.description.abstractWe propose k-SACK, a TCP variant that has considerably improved throughput characteristics over lossy links. A k-SACK source does not consider every packet loss as an indication of congestion. It uses the selective acknowledgement option to estimate a parameter lookahead-loss, which is used in congestion detection. Our results demonstrate that k-SACK maintains a steady performance in a non-lossy environment while showing considerable performance improvement over paths spanning lossy links. Throughput gains of more than 90% are observed over links with packet loss rates of the order of 5%. We show that when k-SACK is used over wireline Internet links, it does not harm the existing TCP variants significantly. k-SACK is simple to implement and does not involve any additional overheadsen
dc.format.extent160422 bytes-
dc.format.mimetypeapplication/msword-
dc.language.isoenen
dc.subjectTCPen
dc.subjectSelective Acknowledgements (SACK)en
dc.subjectWireless networksen
dc.subjectLossy linksen
dc.subjectBandwidth utilizationen
dc.subjectThroughputen
dc.subjectFairnessen
dc.subjectDrop tailen
dc.subjectCongestionen
dc.titleTCP k-SACK: a simple protocol to improve performance over lossy linksen
dc.typeArticleen
Appears in Collections:Computer Science and Engineering

Files in This Item:

File Description SizeFormat
chrungootcp2001.pdf171.63 kBAdobe PDFView/Open
View Statistics

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

 

Valid XHTML 1.0! DSpace Software Copyright © 2002-2010  Duraspace - Feedback