DocumentCode
2502798
Title
A demonstration of the FN packet marking probability
Author
Kadhum, Mohammed M. ; Hassan, Suhaidi
Author_Institution
InterNetWorks Res. Group, Univ. Utara Malaysia, Sintok, Malaysia
fYear
2009
fDate
28-30 Sept. 2009
Firstpage
1540
Lastpage
1543
Abstract
The effectiveness of queue management mechanisms relies on how good their control decisions will help in satisfying their goals regarding congestion avoidance and control. These decisions are implemented and compelled during the design of the packet mark probability and the mark activation functions. The design of fast congestion notification (FN) drop/mark probability function enables the two control decisions, packet admissions and congestion control directing, to be made along with each other. This permits sending congestion avoidance notification as early as required even if the queue is almost empty, and preventing congestion notification even if the queue is almost full but the arrival rate is controllable. This leads to good buffer utilization and proper congestion detection. This paper demonstrates the drop/mark probability functions that the fast FN policy would exercise for different values of optimal queue size, and also for a specific optimal queue size value.
Keywords
decision theory; probability; queueing theory; telecommunication congestion control; FN packet marking probability; arrival rate control; buffer utilization; congestion avoidance; drop-mark probability function; queue management mechanism; Art; Buffer overflow; Delay; Educational institutions; Internet; Optimal control; Probability; Quality of service; Throughput; Traffic control; Fast congestion Notification (FN) gateway; Packet drop probability; Random Early Detection (RED) gateway;
fLanguage
English
Publisher
ieee
Conference_Titel
Communications and Information Technology, 2009. ISCIT 2009. 9th International Symposium on
Conference_Location
Icheon
Print_ISBN
978-1-4244-4521-9
Electronic_ISBN
978-1-4244-4522-6
Type
conf
DOI
10.1109/ISCIT.2009.5341027
Filename
5341027
Link To Document