Hierarchical Multicast Protocols with Quality of Service

Abstract

Multimedia applications vary radically from the traditional applications of data transfers such as email or file transfer. Indeed, these applications generally concern only two users: a source and a destination. In addition, the transmission delays do not influence the service. On the contrary, multimedia applications can imply more than two users (a videoconference for 100 people for example). In addition, these applications need short delays and flow guarantees in order to ensure a continuous playback of the multimedia flow. However, the current networks generally provide only a service, which is "at best" and point-to-point. Particular protocols and mechanisms must be developed at both the transport and application levels. Some mechanisms already exist: they enable the applications, using the multicast over Internet, to operate generally in satisfactory conditions. Hence, it becomes necessary to provide guarantees in terms of quality of service for this type of multimedia application.The quality of service (QoS) refers to the way a packet is delivered. Thus, it is defined by the following parameters:-the delay, which characterizes the end-to-end transfer time; -the jitter, which represents the variation of the communication delay;Chapter written by Abderrahim BENSLIMANE and Omar MOUSSAOUI.52 Multimedia Multicast on the Internet -the bandwidth, which corresponds to the possible throughput between two end entities; it is limited by the throughput of traveled physical links, but also by the concurrent flows and the equipment capacity; -the reliability, which is the average error ratio of various communication supports and equipment.Several studies were suggested in order to guarantee a certain QoS [BRA 97], or at least to differentiate the services [BLA 98]. However, there is still a lot left to be done in order to deploy the multicast protocols while taking into account QoS.Multicast routing algorithms and protocols have been largely studied in the literature [BAL 97, BIS 00, DEE 99, DID 97, EST 95, EST 98]. However, most of them are not scalable and do not manage the QoS efficiently. In addition, there were several studies advocating the importance of QoS in multicast routing [CAR 97, CHE 00, FAL 98, SRI 98]. These works mainly deal with the research of paths from the new members towards the multicast tree by considering the parameters of QoS, which is done by taking into account either a single path or multiple paths. However, with a single path, QoS is not necessarily considered.On the other hand, with multiple paths, there is an overloading of the network (i.e. reservation cost, flooding), which does not enable scalability. The development of multicast routing sensitive to QoS drew less attention, even though it is indispensable for multimedia applications. As for the scalability of multicast routing protocols, the solutions based on "hierarchical trees" are very promising [THY 95, HOF 96, PRA 01]. These protocols decompose the global multicast group into separated sub-groups. Each group consists of partic...

A Scalable Multicast Protocol with QoS Guaranties

“…It is bidirectional because there is no distinction between sources and receivers. The core based tree (CBT) [2] and PIM Sparce Mode (PIM-SM) [7] are well known examples of shared tree protocol. Bach sender transmits messages to the multicast group toward the core.…”

“…The first one is called RDV (RenDezVous) hierarchical architecture and the second LDS (Link Direct Shortest path) hierarchical architecture. Complete description and algorithms of these two methods can be found in [2].…”

A PBIL for Load Balancing in Network Coding Based Multicasting

“…100 Mbps. The consumed bandwidth of link e i 2 E prior to NCM is randomly generated in the range [1,50] Mbps with a uniform distribution. The bandwidth consumption of each link in NCM subgraph is set to 30 Mbps.…”

“…Multicast is one of the efficient technologies developed for supporting one-to-many multimedia applications with stringent quality-of-service (QoS). Therefore, this technology has drawn significant amount of research attention from both academia and industry [1]. Unfortunately, multicast with store-and-forward strategy cannot guarantee that a theoretical maximal throughput is always obtained.…”

A quantum inspired evolutionary algorithm for dynamic multicast routing with network coding

“…Network coding is featured with a number of outstanding advantages, such as natural load balancing, robustness against network failures, stable resistance to attacks, energy saving in multi-hop wireless networks, and so on [1]. Multicast is a technology for point to multi-point data transmission, where the source expects to deliver identical data to a set of receivers [2]. Network coding, when applied to multicast, can always enable the achievement of theoretically maximized throughput [3].…”