DocumentCode :
3537304
Title :
A deadline-aware rate-adaptive packet pacing algorithm for small buffer networks
Author :
Venkatesh, T. ; Prakash, K. Phani
Author_Institution :
Dept. of Comput. Sci. & Eng., Indian Inst. of Technol. Guwahati, Guwahati, India
fYear :
2011
fDate :
14-16 Dec. 2011
Firstpage :
189
Lastpage :
194
Abstract :
Recent research on sizing router buffers in high speed networks promises the realization of core network with buffer enough to store just a few packets. However, such networks work well with traffic efficiently paced to reduce both short-term and long-range burstiness. In this work, we propose an efficient pacing scheme at the edge network that works for packets with or without pre-determined deadlines and smoothen the traffic according to the occupancy of the router buffer. The proposed algorithm paces the traffic both according to the delay and queue length constraints. It is thus shown to possess the advantages of both delay-based pacing and queue length-based pacing algorithms in the literature and yet overcome their shortcomings. We demonstrate through simulations that the proposed pacing algorithm significantly reduces the loss rate even when the traffic contains significant UDP traffic compared to the best known pacing schemes in the literature.
Keywords :
delays; queueing theory; telecommunication network routing; telecommunication traffic; UDP traffic; core network; deadline-aware rate-adaptive packet pacing algorithm; delay constraints; delay-based pacing algorithms; edge network; high speed networks; queue length constraints; queue length-based pacing algorithms; sizing router buffers; small buffer networks; Algorithm design and analysis; Buffer storage; Compaction; Delay; Loss measurement; Optical buffering; Optical losses;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Networks (ICON), 2011 17th IEEE International Conference on
Conference_Location :
Singapore
ISSN :
1556-6463
Print_ISBN :
978-1-4577-1824-3
Type :
conf
DOI :
10.1109/ICON.2011.6168473
Filename :
6168473
Link To Document :
بازگشت