On Exploiting Traffic Predictability in Active Queue Management
نویسندگان
چکیده
Analytical and empirical studies have shown that self-similar trafic can have detrimental impact on network performance including amplified queuing delay and packet loss ratio. On the flip side, the ubiquity of scale-invariant burstiness observed across diverse networking contexts can be exploited to better design resource control algorithms. In this paper, we explore the issue of exploiting traflc predictability to enhance the performance of active queue management (AQM). We show that the correlation structure present in long-range dependent tra@c can be detected on-line and used to accurately predict the future tra@c. We then design, with the objective of stabling the instantaneous queue length at a desirable level, a LMMSE-based controller, and figure in the prediction results in the calculation of the packet dropping probability. The resulting scheme is termed as predictive AQM (PAQM). Through analytical reasoning, we show that PAQM is a generalized version of RED with a new dimension of congestion index ~ the amount of trafic that will arrive in the next few measurement intervals. By stabilizing the queue at a desirable level with consideration of f&are trafic, PAQM enables the link capacity to be fally utilized, while not incurring excessive packet loss. Through ns-2 simulation, we compare PAQM against existing AQM schemes with respect to different performance criteria. In particular, we show that under most cases PAQM outperforms SRED in stabilizing the instantaneous queue length, and adaptive virtual queue (AVQ) in reducing packet loss ratio and better utilizing the link capacity.
منابع مشابه
On the Predictability of Data Network Traffic
The predictably of data network traffic is assessed. Different topologies, types of traffic, and queueing disciplines are studied. Linear and nonlinear AR(MA) models as well as state space, and models based on canonical correlation are employed. These predictors are compared against two simple predictors: 1. the prediction is the mean value of the time series, 2. the prediction is the last obse...
متن کاملEvaluation of Active Queue Management Algorithms
Active Queue Management (AQM) is a very active research area in networking flows. In order to stem the increasing packet loss rates caused by an exponential increase in network traffic, researchers have been considering the deployment of active queue management algorithms. In this paper we will evaluate the active queue management algorithms such as Droptail, RED, BLUE, REM, GREEN and PURPLE. T...
متن کاملIngress Traffic Control in Differentiated Services Ip Networks
xii Acknowledgements xiv 1 Traffic Control 1 1.1 About Terminology . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Traffic Control in Circuit-Switched Networks . . . . . . . . . . 3 1.3 Traffic Control in Differentiated Services Networks . . . . . . . 5 1.3.1 Provisioning . . . . . . . . . . . . . . . . . . . . . . . . 7 1.3.2 Management . . . . . . . . . . . . . . . . . . . . . . . . 8 1...
متن کاملTitle of Dissertation : MULTIFRACTAL INTERNET TRAFFIC MODEL AND ACTIVE QUEUE MANAGEMENT
Title of Dissertation: MULTIFRACTAL INTERNET TRAFFIC MODEL AND ACTIVE QUEUE MANAGEMENT Jia-Shiang Jou, Doctor of Philosophy, 2003 Dissertation directed by: Professor John S. Baras Department of Electrical and Computer Engineering We propose a multilevel (hierarchical) ON/OFF model to simultaneously capture the mono/multifractal behavior of Internet traffic. Parameter estimation methods are deve...
متن کامل