Your Search Results

Use this resource - and many more! - in your textbook!

AcademicPub holds over eight million pieces of educational content for you to mix-and-match your way.

Experience the freedom of customizing your course pack with AcademicPub!
Not an educator but still interested in using this content? No problem! Visit our provider's page to contact the publisher and get permission directly.

Using auto-tuning proportional integral probability to improve random early detection

By: Hao Wang; Bo Wang; Zilong Ye;

2011 / IEEE / 978-1-61284-307-0

Description

This item was taken from the IEEE Conference ' Using auto-tuning proportional integral probability to improve random early detection ' Random early detection (RED), a well-known active queue management (AQM) scheme, has been popularly deployed by many router vendors. But RED is very sensitive to traffic load and parameter configuration, and its equilibrium queue length varies greatly with the congestion degree and parameter settings. To solve the above problems, this paper proposes improved RED (named IRED) by using auto-tuning proportional integral (PI) probability. An adaptation mechanism is designed to adjust the maximum packet marking probability for stable average queue length. The key concept is that when the traffic load changes and the queue length deviates from the target value, we adjust the maximum packet marking probability to drive the queue length to the target, which meets the goal of AQM design. Extensive simulations are conducted to verify the validity of IRED. The results confirm that IRED is superior to RED and its variant in terms of stability and robustness. IRED is not sensitive to traffic loads and can maintain stable average queue length in spite of congestion degrees. In addition, IRED makes very few changes to the RED algorithm, and overcomes the RED's shortcomings without introducing extra variable or much overhead.