Table of Contents
Characterization of Internet traffic and its impact on network performance
Who we are: the university
Who we are: our laboratory
Who we are: our laboratory
Characterization of Internet traffic
Introduction: traffic patterns
Introduction: simulations
Requirements for data transmission
Quality of Service parameters
Traffic patterns: “Star Wars” trace
Definitions: packet loss and loss episodes
Contribution of loss episodes
Collection of packet loss data
ns-2 network simulator
Simple simulation scenario
Complex simulation scenario
Visualization: textured dot strip plots
Contribution of packet loss
Packet loss rates:mixed UDP and TCP sources
Gilbert model
Comparison of simulation datawith the Gilbert model
Heavy-tailed distribution of packet loss
Wavelet analysis
Filter-bank pyramidal algorithm
LRD and Hurst parameter
Wavelet-based Hurst parameter estimator
Abry-Veitch estimator
Self-similar traffic patterns
Influence of time-scales
Influence of time-scales
PPT Slide
PPT Slide
PPT Slide
PPT Slide
PPT Slide
Characteristics ofthe “Star Wars” traffic trace
Wavelet analysis of packet loss
PPT Slide
Analysis of packet loss on multiple time-scales
Packet loss and time-scales
Hurst parameter estimation
Concluding remarks (1)
Concluding remarks (2)
Concluding remarks (3)
Current work
Collaborators
References: on traffic
References: on packet loss
Vancouver
Vancouver
Vancouver
|