A Comparison of Application-Level and Router-Assisted Hierarchical Schemes for Reliable Multicast

Radoslavov, Pavlin; Papadopoulos, Christos; Govindan, Ramesh; Estrin, Deborah

doi:10.1109/tnet.2004.828950

Cited by 24 publications

(8 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is shown by Radoslavov et al (2004) that application-level and router-assisted schemes for reliable multicast have similar performances, which is counter-intuitive but conforms the idea that multicast reliability should be implemented at end-hosts only. Centralised and distributed error recovery approaches are compared in terms of bandwidth and latency by Lacher et al (2000) with the conclusion that distributed approaches perform better in terms of bandwidth but centralised approaches are more desirable because they are easier to deploy and do not require multicast retransmission capability.…”

Section: Related Workmentioning

confidence: 62%

Theoretical analysis and comparison of various approaches for reliable multicast

Nadeau

Anjali

2007

IJITST

View full text Add to dashboard Cite

Scalability is always an issue for multicast protocols. This paper analyses and compares eight protocols for packet loss recovery in multicast communications. The goal of our analysis is to emphasise which protocols provide good recovery times without inducing too much extra traffic in the network, which is the limiting factor for scalability. Assuming the same loss rates on every link in the network, we make a statistical analysis to derive the expected number of packet retransmissions and compare the performance of the protocols. The results show great variations between protocols and provide good insights for the design of new protocols.

show abstract

Section: Related Workmentioning

confidence: 62%

Theoretical analysis and comparison of various approaches for reliable multicast

Nadeau

Anjali

2007

IJITST

View full text Add to dashboard Cite

show abstract

“…If any of the sinks were not able to decode the generation they signal that they needs additional information which the source sends. This approach adapts better to changing channel conditions and as such can utilise the channel better, however, the feedback from the sinks introduces the exposure problem [16] and the crying baby problem [17]. Thus this approach works best if the sinks have relatively uniform channel conditions, and if the number of sinks is moderate.…”

Section: Protocol Considerationsmentioning

confidence: 99%

“…Thus an interesting approach is to let the sinks cooperate and thus exploit the connection diversity. Instead of a sink requesting additionally data specifically from the source, any node that received the request could respond, thus more than one node could potentially attempt to answer the request, which would introduce the implosion problem [16]. One of the main drawbacks of this approach is the high complexity it introduces, one technique to remedy this could be the NC.…”

Section: Protocol Considerationsmentioning

confidence: 99%

A mobile application prototype using network coding

Pedersen

Heide

Fitzek

et al. 2010

Trans Emerging Tel Tech

View full text Add to dashboard Cite

SUMMARYThis paper looks into implementation details of network coding for a mobile application running on commercial mobile phones. We describe the necessary coding operations and algorithms that implements them. The coding algorithms forms the basis for a implementation in C++ and Symbian C++. We report on practical measurement results of coding throughput and energy consumption for a single-source multiple-sinks network, with and without recoding at the sinks. These results confirm that network coding is practical even on computationally weak platforms, and that network coding potentially can be used to reduce energy consumption.

show abstract

“…More recent work includes a comparison study of the costs of application-layer reliable multicast schemes and router-assisted schemes [42]. Finally, He [43] presents a comparison study between incremental deployment of LMS and PGM.…”

Section: Network Assisted Schemesmentioning

confidence: 99%

Light-Weight Multicast Services (LMS): A Router-Assisted Scheme for Reliable Multicast

Papadopoulos

Parulkar

Varghese

2004

IEEE/ACM Trans. Networking

Self Cite

View full text Add to dashboard Cite

--Building on the success of unicast IP, IP IntroductionAt the core of the Internet architecture lies the simplicity and elegance of IP and its design principles [11]. Internet architects realized early on that by foregoing the wire-like robustness of traditional communications networks (such as the telephone network) and pushing the intelligence to the edges, a network can be built on a much simpler, cheaper and highly scalable infrastructure.The resulting best-effort service model has proven highly flexible. However, the interaction of store-and-forward packet forwarding, finite buffers and bursty sources occasionally leads to congestion and loss. Applications requiring better than best-effort reliability must counteract loss with error control, the component of a communication protocol responsible for reliability [1], [2].IP Multicast [5] debuted in the late 1980s and was hailed as a natural extension of the unicast model. Multicast is a powerful service because it allows a single source to reach a virtually unlimited number of receivers in a very efficient and scalable manner. Multicast is well-suited for applications such as streaming media, distance learning, Internet radio and television, distributed interactive simulation, file transfer, software updates and much more. Continuing the architectural tradition of unicast, IP Multicast adopted a simple, best-effort, anonymous, broadcast-like service, often compared to a radio dial-tone: anyone may tune in and anyone may transmit. Thus, similar to unicast, IP Multicast provides a general service on top of which richer services can be built.Despite the vigorous promotion of multicast by both the research and industry communities, the Internet Service Providers (ISPs) and the users have not yet embraced the service. Many reasons have been cited [38], [39], which include difficulties with inter-domain routing, peering relationships, address allocation, limited address space, security, billing, and the lack of a general scalable and reliable transport service analogous to TCP. This paper addresses the last issue. We present Lightweight Multicast Services (LMS), an extension to IP Multicast, on top of which a general and scalable reliable multicast transport service can be constructed. LMS extends IP Multicast with a set of simple and lightweight services that enhance router forwarding to enable highly scalable, network assisted solutions to reliable multicast. LMS cleanly separates the transport and forwarding components of error control, keeps the former at the endpoints thus avoiding layer violations, and pushes the latter to the routers where it can be implemented most efficiently.Unicast error control mechanisms are not suitable for large-scale multicast due to the many-to-many nature of IP Multicast. Losses in multicast typically affect part of the multicast tree and attempting to recover localized loss leads to the following problems:• Implosion: occurs when the loss of a packet triggers redundant messages (requests and/or retransmissions). In lar...

show abstract

A Comparison of Application-Level and Router-Assisted Hierarchical Schemes for Reliable Multicast

Cited by 24 publications

References 16 publications

Theoretical analysis and comparison of various approaches for reliable multicast

Theoretical analysis and comparison of various approaches for reliable multicast

A mobile application prototype using network coding

Light-Weight Multicast Services (LMS): A Router-Assisted Scheme for Reliable Multicast

Contact Info

Product

Resources

About