Abstract — Recent multicast routing protocol proposals such as protocol independent multicast (PIM) and core-based trees (CBT) have been based on the notion of group-shared trees. Since construction of a minimal-cost tree spanning all members of a group is difficult, they rely on center-based trees and distribute packets from all sources over a single shortest-path tree rooted at some center. PIM and CBT provisionally use administrative selection or simple heuristics for locating the center of a group but do not preclude the use of other methods that provide an ordered list of centers. Other previously proposed heuristics typically require knowledge of the complete network topology, a requirement which is not always practical for a distributed problem such as Internet routing. In this paper we investigate the problem of finding a good center in distributed fashion, study various heuristics for automating center selection, and examine their applicability to real-world networks. We also propose several new algorithms which we feel to be more practical than existing methods. We present simulation results on hierarchical and nonhierarchical networks showing that of the methods potentially feasible in the Internet multicast backbone, ours offer the best results in terms of cost and delay, and they incur low overhead.

Multicasting is a technique that enables a single packet transmission to reach one or more destinations or group. The primary benefits of a packet reaching multiple destinations from a single transmission are threefold: bandwidth minimization; the exploitation of parallelism in the network; the optimization of transmitter costs. In this thesis we investigate and analyse each of the different network layer multicast algorithms and protocols, looking in particular at their scalability, since multicast scalability was the primary motivator for this work. Our first and most significant contribution involves the presentation of a new multicast architecture and protocol, designed for best-effort, connectionless datagram networks such as the IP Internet. This new architecture typically offers considerably more favourable scaling characteristics than do existing multicast schemes. Our other most significant contribution is the security architecture that is integral in our new multicast proposa...

The Core-based approach is inevitable in multicast routing protocols as it provides efficient management of multicast path in changing group memberships, and scalability and performance. In this paper, we present a comprehensive analysis of this approach with the emphasis on core selection for the first time in literature. We first examine the evolution of multicast routing protocols into the core-based architecture and the motivation for the approach. Then we review the core-selection algorithms in the literature for their algorithmic structure and performance. Our study involves an extensive computational and message complexity analysis of each algorithm, and a classification for their deployment characteristics and algorithmic complexities. To the best of our knowledge ours is the first paper providing such extensive comparative analysis of core-based multicast routing protocols. 1.

MC routing protocols currently being standardized, PIM and CBT, propose the use of shared center-based trees, but they do not specify how to select the centers of these trees. We survey previous work on the center selection problems in the fields of operations research and communication networks. 1

We study the problem of constructing delay-constrained shared multicast trees for realtime applications in connection-oriented high-speed networks. We formulate the problem as a diameter-constrained minimum Steiner tree problem, and prove that it is NP-complete. Then we propose distributed, dynamic heuristics for solving this problem. Most previous work on shared multicast trees starts by selecting a multicast center, and then constructs a shared tree around it. One of the heuristics we propose eliminates the need for a dedicated center selection phase. We use simulation to evaluate the performance of the proposed heuristics. # This work was supported in part by the Center for Advanced Computing and Communication at North Carolina State University, and by AFOSR grants F49620-92-J-0441DEF and F49620-96-1-0061. 1 1 Introduction Broadband Integrated Services Digital Networks (B-ISDN) [1] are evolving at a fast pace. B-ISDNs use the Asynchronous Transfer Mode (ATM) [1] which is ...

In past years, researchers have proposed the Core Based Tree (CBT) and Protocol Independent Multicast (PIM) protocols to route multicast data on the Internet. Such protocols need to locate the Core of a group to achieve efficient multicast routing. In this paper, we propose a scalable distributed protocol that can be used to move the Core to near-optimal location in the dynamic multicast tree, and that allows the Core to migrate efficiently when the multicast tree is expanded or shrunk. Our protocol does not require knowledge of the complete network topology, and information of all the members is distributed among local Agents; the Core only maintains the information of Agents of the group. Also, only the Agents participate in Core selection. Therefore, the proposed protocol reduces the runtime overhead and message complexity while performing Core migration.

Many distributed multimedia applications involve data delivery from a source to multiple destinations, the participating nodes forming a multicast group. In the naive solution, a separate connection can be established from the source to other group nodes. However, better solutions have been devised. A tree can be established for each source with the participants as the leaf nodes or just have one tree spanning all the participants. There are several advantages of having a single tree for the multicast group. In this paper, we introduce a model for supporting shared multicast trees over the ATM networks when the number of sources are bounded. Our work will allow the proposed wide area multicast protocols (like CBT and PIM) which allow leaf node initiated joins to be used on ATM networks. 1 Introduction The increasing computing power of workstations and speed of computer networks are ushering in a new era of applications involving multimedia data[12, 17]. Examples include applications ...

The Core-based approach provides efficient management of multicast paths in dynamic environments, scalability and performance in multicast routing protocols. In this paper, we present a comprehensive analysis of this approach with emphasis on core selection. We review core-selection algorithms in the literature for their algorithmic structure and performance. Our study also involves an extensive computational and message complexity analysis of each algorithm, and a classification for their deployment characteristics and algorithmic complexities. To the best of our knowledge ours is the first paper providing such extensive comparative analysis of core-based multicast routing protocols.

Multicast routing protocols are designed to construct and maintain trees. These trees distribute data on a one-to-many basis in which the replication and distribution of data is accomplished by nodes of the tree. Initial efforts in designing the first IP multicast routing protocol focused on the development of algorithms that produced a separate tree for each source of the multicast group. For those edges of the tree that do not have group members, special control messages, referred to as Prunes, are sent hop-by-hop towards the source to ensure that data is only sent to those nodes that have downstream receivers. However, even with the inclusion of Prune messages, each router in the network was required to retain source/group state for every active source of every group. Shared tree algorithms and protocols were developed in an attempt to reduce the amount of state stored by source tree protocols. The fundamental design in this approach centered on leaf routers sending an explicit unicast join message to the shared root of the tree, thus grafting a branch from the receiver onto the shared tree. Two of the more widely known attempts to achieve this are known as Core Based Trees (CBT) and Protocol Independent Multicast (PIM). In general terms,