On Dynamic Multicast Routing Trees in Computer Networks

博士 === 國立清華大學 === 資訊工程學系 === 88 === Multicast routing trees can be classified into two categories, namely, the source-rooted trees and the shared trees. To reduce the cost of a multicast tree for a multicast group with one source node, usually a source-rooted multicast routing tree is established. F...

Full description

Bibliographic Details
Main Authors: Shou-Chuan Lai, 賴守全
Other Authors: Hwa-Chun Lin
Format: Others
Language:en_US
Published: 2000
Online Access:http://ndltd.ncl.edu.tw/handle/53753396948944702290
Description
Summary:博士 === 國立清華大學 === 資訊工程學系 === 88 === Multicast routing trees can be classified into two categories, namely, the source-rooted trees and the shared trees. To reduce the cost of a multicast tree for a multicast group with one source node, usually a source-rooted multicast routing tree is established. For some ap-plications, there may be more than one source node in a multicast group. Algorithms which establish source-rooted trees will have to establish one source-rooted tree for each of the source nodes. Only one shared tree is established for all source nodes in a multicast group if algorithms which establish shared trees are used. In this dissertation, both categories of multicast routing trees are studied. First of all, a new dynamic multicast routing algorithm called Virtual Trunk Dynamic Multicast (VTDM) routing algorithm is proposed to establish low-cost source-rooted mul-ticast routing trees for multicast connections which allow nodes to join and leave the multicast connections dynamically. Next, the problem of core placement for the Core Based Tree (CBT) multicast routing architecture is studied. The CBT multicast routing architecture, one of the promising multicast routing protocols for the Internet, establishes shared trees for multicast connections. The center node of the shared tree is called the core. The cost of the shared tree can be reduced if the location of the core is properly selected. The perfor-mance of dynamic multicast routing algorithms are usually studied by generating random graphs to represent different networks and performing extensive simulations based on the random graphs. Depending on the techniques for generating random graphs, the generated random graphs may not resemble the topologies of real IP networks. In order to have better knowledge of the topologies of IP networks, an automatic topology discovery algorithm for IP networks is proposed. The proposed algorithm has been successfully implemented on the Tcl/Tk and Scotty environment.