Open AccessARCH: A dual‐mode fairness‐driven AQM for promoting cooperative behaviour in best effort Internet
Author(s) -
Abbas Ghulam,
Raza Usman,
Halim Zahid,
Kifayat Kashif
Publication year - 2019
Publication title -
iet networks
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.466
H-Index - 21
eISSN - 2047-4962
pISSN - 2047-4954
DOI - 10.1049/iet-net.2018.5089
Subject(s) - active queue management , computer science , computer network , network congestion , stateless protocol , bandwidth (computing) , incentive , queue , fairness measure , network packet , telecommunications , microeconomics , economics , throughput , wireless
The cooperative behaviour of end‐systems in efficiently and fairly utilizing the scarce bandwidth is of paramount significance in the best effort Internet. Not only does the uncooperative behaviour result in unfairness, but it may also induce congestion collapse. How to urge end‐systems to share the bandwidth cooperatively is a problem that depends, in essence, on incentives. This study presents a novel fairness‐driven active queue management (AQM)‐based incentive scheme to detect and take punitive action against any uncooperative behaviour. Different from all other fairness‐driven AQM schemes, the proposed scheme offers a dual‐mode control function that can adaptively switch between stateless and partial‐state modes in order to be scalable. The control function remains stateless at low to moderate congestion levels and maintains a partial‐state when congestion becomes severe. In the stateless mode, the identification and restriction of unresponsive flows are based on a matched‐drop mechanism having a dynamic drawing factor to differentially drop packets from unresponsive flows. In the partial‐state mode, a fair‐share is estimated and flows whose arrival rates exceed the fair‐share bandwidth get their rates limited to the fair rate. Simulation results demonstrate the reasonable performance improvement of the proposed scheme over well‐known fairness‐driven AQM schemes.