The performance of QoS-aware IP multicast routing protocols

Research in the area of QoS‐aware dynamic multicast routing protocols has been very active in recent years. Protocols based on dynamic Steiner tree strategies, such as YAM and QoSMIC, have been consistently shown to outperform those based on shortest path heuristics, such as PIM and DVMRP. However,...

Full description

Saved in:
Bibliographic Details
Published inNetworks Vol. 42; no. 2; pp. 97 - 108
Main Authors Tseng, Chih-Jen, Chen, Chyou-Hwa
Format Journal Article
LanguageEnglish
Published Hoboken Wiley Subscription Services, Inc., A Wiley Company 01.09.2003
John Wiley & Sons
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:Research in the area of QoS‐aware dynamic multicast routing protocols has been very active in recent years. Protocols based on dynamic Steiner tree strategies, such as YAM and QoSMIC, have been consistently shown to outperform those based on shortest path heuristics, such as PIM and DVMRP. However, these protocols all suffer from the problem of poor scalability for one or more of the following reasons: high control overhead, insufficient robustness with the adoption of a centralized group manager, and excessively long join latency. In addition, these protocols perform well only when group members are either densely populated or sparsely populated, but, unfortunately, not both. In this paper, we propose a protocol, named DSDMR, which can adapt its strategy based on sensed group member densities. Underlying DSDMR is an adaptive two‐direction join mechanism that tries to find good attaching points for new group members either from the source or from the new joining member depending on member densities. We evaluate our scheme using extensive simulations and found that DSDMR can build multicast trees with costs close to the best greedy strategy, very low control overhead, and very short join latency across a wide member density spectrum. Furthermore, its success ratio is only slightly lower than is the best greedy strategy in finding feasible routes subject to both bandwidth and end‐to‐end delay constraints. © 2003 Wiley Periodicals, Inc.
Bibliography:istex:360DBDEEAB32050F346036EBBED453340F2B0CE5
An extended version of the paper was published in the Proceedings of 15th IEEE International Conference on Information Networking.
ArticleID:NET10084
ark:/67375/WNG-3S9W7QR0-B
ISSN:0028-3045
1097-0037
DOI:10.1002/net.10084