返回Google圖書搜尋

Reducing Packet-loss by Taking Long-range Dependences Into Account

J. Ignacio Alvarez-HamelinPierre Fraigniaud

出版	Université de Paris-Sud, Centre d'Orsay, Laboratoire de Recherche en Informatique, 2002
URL	http://books.google.com.hk/books?id=2_Y7HAAACAAJ&hl=&source=gbs_api

註釋Abstract: "We show that the 'fractal' behavior of Internet traffic can be efficiently and practically employed to significantly reduce packet-loss. Thanks to recent advances in the theory of self-similar processes, we define the probabilistic congestion of a link, based on an estimated computation of the packet-loss probability over that link. This congestion parameter captures qualitative characteristics of the traffic, which are shown to be much more stable than usual quantitative characteristics such as current delay or available bandwidth. Hence, it allows valid predictions on the future behavior of the network, on which one can base efficient routing strategies. We show how to implement the computation of the probabilistic congestion, and we illustrate one of its simple applications for improving multicast protocols like QoSMIC or YAM. Simulations based on real traffic samples demonstrate significant reductions of packet loss in QoSMIC using probabilistic congestion, compared to QoSMIC using standard QoS criteria. Unicast routing should also benefit from the use of the probabilistic congestion while preserving the best effort nature of IP, i.e., without resource reservation."