Title :
Advances in TCP congestion prevention
Author :
Hruby, M. ; Olsovsky, M. ; Kotocova, M.
Author_Institution :
Inst. of Comput. Syst. & Networks, Slovak Univ. of Technol. in Bratislava, Bratislava, Slovakia
Abstract :
Today´s congestion prevention in IP networks is based mainly on hard TCP end nodes notification. This notification uses queuing mechanism and ends up in certain packet drop which forces specific TCP end nodes to decrease the congestion window. Different way is soft TCP end nodes notification which is based on existing ECN system while allowing network to inform all TCP end nodes to decrease the congestion window. Both ways are usable but do not distinguish concurrent TCP flows while notifying end nodes. This leads to network performance issues and degradation when important TCP flows may not be recognized and will be suffering from starvation caused by old and greedy TCP flows. This paper introduces a new way of increasing network performance called ACNS by means of modifying the most common transport protocol TCP. ACNS itself is focused on congestion prevention by means of penalizing old flows and providing more accurate details for loss type classification.
Keywords :
Internet; queueing theory; telecommunication congestion control; transport protocols; ACNS network performance; ECN system; IP network; Internet protocol; TCP congestion prevention; advanced congestion notification system; concurrent TCP flow; congestion window; explicit congestion notification; hard TCP end node notification; network degradation; network performance; packet drop; queuing mechanism; soft TCP end node notification; transport control protocol; Fluid flow measurement; IP networks; Informatics; Machine intelligence; Market research; Mathematical model; Standards; TCP; age; congestion control; flow trend; loss rate; network; notification system; performance;
Conference_Titel :
Applied Machine Intelligence and Informatics (SAMI), 2013 IEEE 11th International Symposium on
Conference_Location :
Herl´any
Print_ISBN :
978-1-4673-5928-3
Electronic_ISBN :
978-1-4673-5927-6
DOI :
10.1109/SAMI.2013.6480956