Bandwidth Trading and Management in VPN Service Overlay Network

碩士 === 國立臺灣大學 === 資訊管理研究所 === 91 ===   Since businesses started to use computers in more than one location, there have been the requirements to connect them together in a secure and private way to facilitate corporate communication. Virtual Private Network (VPN) is proposed to meet the re...

Full description

Bibliographic Details
Main Authors: Wu, Liang-Ming, 武良銘
Other Authors: Yeali S. Sun
Format: Others
Language:zh-TW
Published: 2003
Online Access:http://ndltd.ncl.edu.tw/handle/76140167282723618621
Description
Summary:碩士 === 國立臺灣大學 === 資訊管理研究所 === 91 ===   Since businesses started to use computers in more than one location, there have been the requirements to connect them together in a secure and private way to facilitate corporate communication. Virtual Private Network (VPN) is proposed to meet the requirements where the architecture and relevant technologies of VPN have evolved during the past decades and have became mature today.   However, to provide guaranteed quality of service for a VPN is not extensively concentrated as well as secure issues. In this thesis, we propose a new service model on a 3-layered overlay network that focuses on providing quality of service for a VPN.   We propose the concept of bandwidth trading and sequentially define a bandwidth trading service model that if a VPN’s original reserved bandwidth is insufficient, trading events will occur therefore it will borrow the insufficiency from others who have residual bandwidth to lend. Through temporary borrowing, not only the VPN’s blocking rate and packets loss number efficiently reduce, but the utilization of bandwidth resource of operational network efficiently rises also. In this service model we also take the fairness of trading into account, in addition, to achieve adaptive traffic engineering is one of the goals. By means of the simulation experiments, we prove that the service model can effectively attain the goals. In the other hands, the experiments also show that the feasibility of the service model.