Wavelet Based Estimators of Long-Range Dependencies in Traffic Traces

2/25/03


Click here to start


Table of Contents

Wavelet Based Estimators of Long-Range Dependencies in Traffic Traces

Road map

Introduction: traffic patterns

Influence of time-scales

Influence of time-scales

LRD and Hurst parameter

LRD and Hurst parameter

Wavelet analysis

Wavelets and multiresolution analysis

Discreet wavelet transformation

Wavelet analysis

Filter-bank pyramidal algorithm

Daubechies wavelet (genus 2)

Daubechies wavelet (genus 9)

Estimation using wavelets

Monofractal estimators

Monofractal estimator

Monofractal estimator of the Hurst parameter

Multifractal estimator

Multifractal estimator

Video traces

MPEG traces

MPEG traces

MPEG traces

Monofractal wavelet estimator: MTV

Multifractal wavelet estimator MTV

Comparison of estimators: MTV

Monofractal wavelet estimator: Mr. Bean

Multifractal wavelet estimator: Mr. Bean

Comparison of estimators: Mr. Bean

Wavelet estimators: performance

Video traces: coders

Video coders: comparison

Wavelet analysis: Futurama (MPEG4)

Wavelet analysis: Futurama (H263)

Comparison of coders: Futurama

Effect of number of vanishing moments: monofractal case

Effect of number of vanishing moments: monofractal case

Effect of vanishing moments: Mr. Bean, monofractal, n = 2

Effect of vanishing moments: Mr. Bean, monofractal, n = 7

Effect of number of vanishing moments: multifractal case

Effect of number of vanishing moments: multifractal case

Effect of vanishing moments: Lambs, multifractal, n = 2

Effect of vanishing moments: Lambs, multifractal, n = 10

Effect of qth-observed moment: multifractal case

Effect of qth-observed moment: multifractal case

Effect of qth-observed moment: News, q = 1

Effect of qth-observed moment: News, q = 10

Wavelet estimators: observations

ns-2 network simulator

Simple simulation scenario

Analysis of loss traces

Wavelet estimator applied to loss processes

Loss trace, 46 packets, monofractal

Loss trace, 46 packets, multifractal

Loss trace, 100 packets, monofractal

Loss trace, 100 packets, multifractal

Loss trace, 200 packets, monofractal

Loss trace, 200 packets, multifractal

Observations: loss processes

Observations: loss processes

Concluding remarks

Concluding remarks

References: on fractal estimators

References: on traffic (1)

References: on traffic (2)

References: on packet loss

References: on queuing schemes

Collaborators

Author: Zelimir Lucic

Email: zlucic@cs.sfu.ca

Home Page: http://www.ensc.sfu.ca/research/cnl/