City University of Hong Kong

CityU Institutional Repository >
3_CityU Electronic Theses and Dissertations >
ETD - Dept. of Manufacturing Engineering and Engineering Management  >
MEEM - Doctor of Philosophy  >

Please use this identifier to cite or link to this item:

Title: Network traffic modeling and queueing analysis based on {DC}-stable self-similar processes
Other Titles: Ji yu {DC} wen tai zi xiang si guo cheng de wang lu liu liang jian mo ji qi pai dui fen xi
基於 {DC} 穩態自相似過程的網路流量建模及其排隊分析
Authors: Ye, Zhipin (葉志頻)
Department: Dept. of Manufacturing Engineering and Engineering Management
Degree: Doctor of Philosophy
Issue Date: 2004
Publisher: City University of Hong Kong
Subjects: Computer networks -- Evaluation
Queuing theory
Telecommunication -- Traffic -- Evaluation
Notes: CityU Call Number: TK5105.5.Y49 2004
Includes bibliographical references (leaves 117-124)
Thesis (Ph.D.)--City University of Hong Kong, 2004
xiii, 125 leaves : ill. ; 30 cm.
Type: Thesis
Abstract: Recent statistical analysis on various types of network traffic has shown that network traffic exhibits two characteristics: self-similar behavior (or long-range dependence) and non-Gaussian heavy-tailed marginal distribution. These new characteristics are not covered by most traditional traffic models which are usually short-range dependent or Gaussian distribution. And traditional queueing theory is inadequate to predict network performance. Therefore, the main objectives in this thesis are to develop a new traffic model in much agreement with the real network traffic data and to present theoretical techniques for queueing analysis and resource allocation. We propose a new a-stable self-similar model based on Linear Fractional Stable Noise for aggregate traffic of modern broadband communication networks. The non-Gaussian model includes a long-range dependent (LRD) time correlation structure and an a-stable marginal distribution. The model is determined by six parameters which describe the burstiness (α), long-range dependence (H, a, b), scale (ơ) and mean value (m) of the traffic. We also present a new method for LRD parameters estimation, which is based on a sample autocorrelation function. Extensive simulation works verify that our model is able to capture the properties of the real traffic. A queueing behavior of a single server with a self-similar input is investigated analytically with the model developed in this thesis. Some scaling properties of queueing length and its distribution are derived, as well as a lower bound for the buffer overflow probability. We also investigate the impact of various parameters of the model on the overflow probability. The parameters are self-similarity H, heaviness α, buffer size x and bandwidth C. It shows that they all have significant effects on network performance. Especially we conclude that the overflow probability is more sensitive with bandwidth than with buffer size. Buffering and multiplexing gain have their most effective and optimal values. Finally, in order to optimize the use of the available bandwidth and buffer, we develop a resource allocation scheme. When effectively facilitated, this can result in significant performance improvements.
Online Catalog Link:
Appears in Collections:MEEM - Doctor of Philosophy

Files in This Item:

File Description SizeFormat
fulltext.html159 BHTMLView/Open
abstract.html159 BHTMLView/Open

Items in CityU IR are protected by copyright, with all rights reserved, unless otherwise indicated.


Valid XHTML 1.0!
DSpace Software © 2013 CityU Library - Send feedback to Library Systems
Privacy Policy · Copyright · Disclaimer