A scalable control topology for multicast communications

Abstract-Application-layer multicast supports group applications without the need for a network-layer multicast protocol. Here, applications arrange themselves in a logical overlay network and transfer data within the overlay. In this paper, we present an application-layer multicast solution that uses a Delaunay triangulation as an overlay network topology. An advantage of using a Delaunay triangulation is that it allows each application to locally derive next-hop routing information without requiring a routing protocol in the overlay. A disadvantage of using a Delaunay triangulation is that the mapping of the overlay to the network topology at the network and data link layer may be suboptimal. We present a protocol, called Delaunay triangulation (DT protocol), which constructs Delaunay triangulation overlay networks. We present measurement experiments of the DT protocol for overlay networks with up to 10 000 members, that are running on a local PC cluster with 100 Linux PCs. The results show that the protocol stabilizes quickly, e.g., an overlay network with 10 000 nodes can be built in just over 30 s. The traffic measurements indicate that the average overhead of a node is only a few kilobits per second if the overlay network is in a steady state. Results of throughput experiments of multicast transmissions (using TCP unicast connections between neighbors in the overlay network) show an achievable throughput of approximately 15 Mb/s in an overlay with 100 nodes and 2 Mb/s in an overlay with 1000 nodes.

show abstract

“…This topology completely ignores the network topology [19]. Data is disseminated using trees which are embedded in the hypercube [21].…”

Section: Comparative Evaluation Of Overlay Topologiesmentioning

confidence: 99%

Application-layer multicast with Delaunay triangulations

Liebeherr

Nahas

GLOBECOM'01. IEEE Global Telecommunications Conference (Cat. No.01CH37270)

Self Cite

View full text Add to dashboard Cite

show abstract

“…A number of research and commercial large-scale parallel machines have been built based on the hypercube network topology [14,22,28,29,31]. Very recently, research has also shown that fault tolerant hypercube networks of large size (e.g., over many thousand nodes) can be used as an effective control topology in supporting large-scale multicast applications in the Internet [19,20].…”

Section: Introductionmentioning

confidence: 99%

Hypercube Network Fault Tolerance: A Probabilistic Approach

2005

View full text Add to dashboard Cite

Extensive experiments and experience have shown that the well-known hypercube networks are highly fault tolerant. What is frustrating is that it seems very difficult to properly formulate and formally prove this important fact, despite extensive research efforts in the past two decades. Most proposed fault tolerance models for hypercube networks are only able to characterize very rare extreme situations thus significantly underestimating the fault tolerance power of hypercube networks, while for more realistic fault tolerance models, the analysis becomes much more complicated. In this paper, we develop new techniques that enable us to analyze a more realistic fault tolerance model and derive lower bounds for the probability of hypercube network fault tolerance in terms of node failure probability. Our results are both theoretically significant and practically important. From the theoretical point of view, our method offers very general and powerful techniques for formally proving lower bounds on the probability of network connectivity, while from the practical point of view, our results provide formally proven and precisely given upper bounds on node failure probabilities for manufacturers to achieve a desired probability for network connectivity. Our techniques are also useful and powerful for analysis of the performance of routing algorithms, and applicable to the study of other hierarchical network structures and to other network communication problems.

show abstract

“…Producing a reliable multicast protocol that scales well with the number of receivers, in terms of network traffic and the processing required of the source and receivers, has proven to be a challenge, as demonstrated by the number of approaches taken in the past (e.g., [5,7,9,11,12,14,15,19,21,20,[27][28][29]). Moreover, as protocols for multicast error control are developed, mechanisms must also be developed for multicast congestion control [3,6,17], similar to those developed for such unicast reliable protocols as TCP [8].…”

Section: Introductionmentioning

confidence: 99%

Organizing multicast receivers deterministically by packet-loss correlation

Levine

Paul²,

Garcia-Luna-Aceves

2003

Multimedia Systems

View full text Add to dashboard Cite

The ability to trace multicast paths is currently available on the Internet by means of Internet Group Management Protocol mtrace packets. We introduce Tracer, a protocol that organizes the receivers of a multicast group deterministically into a logical tree structure while maintaining exact packet-loss correlation for local error recovery, and without requiring any changes to existing multicast routing protocols. Tracer uses mtrace packets in the Internet Group Management Protocol to allow a receiver host to obtain its path to the source of a multicast group. Receivers use the multicast path information to determine how to achieve local error recovery and effective congestion control. We compare the tracing approach with prior mechanisms that attempt local recovery. Results of measurements carried out over the Collaborative Advanced Internet Research Network illustrate the fact that tracing multicast paths is an effective tool to organize receivers based on their packet-loss correlation.

show abstract

A scalable control topology for multicast communications

Cited by 22 publications

References 15 publications

Application-layer multicast with Delaunay triangulations

Application-layer multicast with Delaunay triangulations

Hypercube Network Fault Tolerance: A Probabilistic Approach

Organizing multicast receivers deterministically by packet-loss correlation

Contact Info

Product

Resources

About