Graphs are commonly used to model the topological structure of internetworks, to study problems ranging from routing to resource reservation. A variety of graphs are found in the literature, including fixed topologies such as rings or stars, "well-known" topologies such as the ARPAnet, and randomly generated topologies. While many researchers rely upon graphs for analytic and simulation studies, there has been little analysis of the implications of using a particular model, or how the graph generation method may a ect the results of such studies. Further, the selection of one generation method over another is often arbitrary, since the differences and similarities between methods are not well understood. This paper considers the problem of generating and selecting graph models that reflect the properties of real internetworks. We review generation methods in common use, and also propose several new methods. We consider a set of metrics that characterize the graphs produced by a method, and we quantify similarities and differences amongst several generation methods with respect to these metrics. We also consider the effect of the graph model in the context of a speciffic problem, namely multicast routing.

This paper presents a new protocol, the Ordered Core Based Tree (OCBT) protocol, which remedies several shortcomings of the Core Based Tree (CBT) multicast protocol. We show that the CBT protocol can form loops during periods of routing instability, and that it can consistently fail to build a connected multicast tree, even when the underlying routing is stable. The OCBT protocol provably eliminates these deficiencies and reduces the latency of tree repair following a link or core failure. OCBT also improves scalability by allowing flexible placement of the cores that serve as points of connection to a multicast tree. Simulation results show that the amount of control traffic in OCBT is comparable to that in CBT. 1 Introduction Multicast routing protocols build routing trees for the dissemination of messages to a select group of other stations. In some protocols, such as the Distance Vector Multicast Routing Protocol (DVMRP) [1] and the Protocol Independent Multicast-Dense Mode (PIM-...

vii Acknowledgments viii 1. INTRODUCTION 1 2. BACKGROUND 4 2.1 Looping in CBT : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 5 3. THE OCBT PROTOCOL 11 3.1 Differences between OCBT and CBT : : : : : : : : : : : : : : : : : : : : : : 11 3.2 Core Placement : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 13 3.3 Tree Construction : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 14 3.4 Tree Maintenance : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 18 3.5 Network Partitions : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 19 3.6 OCBT Specification : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 21 4. CORRECTNESS OF OCBT 27 4.1 Connectivity in a Connected Network : : : : : : : : : : : : : : : : : : : : : 27 4.2 Loop Freedom : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 28 4.3 Connectivity in a Partitioned Network : : : : : : : : : : : : : : : : : : : : : 41 5. PERFO...

Graph models are commonly used in studying solutions to internetworking problems. This paper considers several random graph models that have been used to model internetworks, and considers ways to characterize the properties of these graphs. By matching the characteristics of the random graphs to the characteristics of real internetworks, more accurate modeling can be achieved. College of Computing Georgia Institute of Technology Atlanta, Georgia 30332--0280 1 Introduction Accompanying the rapidly increasing use of large networks and internetworks has been a corresponding growth in research on internetworking. Researchers working in internetwork communication generally adopt one of three methods to evaluate their ideas: ffl They implement their ideas on a real internetwork. This approach has obvious difficulties, including the problem of getting access to an internetwork of reasonable size that can tolerate experimentation. ffl They implement their ideas on a smaller network and e...

This paper presents a new protocol, the Ordered Core Based Tree (OCBT) protocol, which remedies several shortcomings of the Core Based Tree (CBT) multicast protocol. We show that the CBT protocol can form loops during periods of routing instability, and that it can consistently fail to build a connected multicast tree, even when the underlying routing is stable. The OCBT protocol provably eliminates these deficiencies and reduces the latency of tree repair following a link or core failure. OCBT also improves scalability by allowing flexible placement of the cores that serve as points of connection to a multicast tree. Simulation results show that the amount of control traffic in OCBT is comparable to that in CBT. 1 Introduction Multicast routing protocols build routing trees for the dissemination of messages to a select group of other stations. In some protocols, such as the Distance Vector Multicast Routing Protocol (DVMRP) [1] and the Protocol Independent Multicast-Dense Mode (PIM-...

this paper we propose an enhancement to the Distance Vector Multicast Routing Protocol (DVMRP), called