An Analytical Model of a P2P Network for Real-Time Multimedia Communications Giuseppe Incarbone, Alfio Lombardo, Giovanni Schembra Dipartimento di Ingegneria Informatica e delle Telecomunicazioni - University of Catania V.le A. Doria, 6 - 95125 Catania - ITALY E-mail: (giuseppe.incarbone, lombardo, schembra)@diit.unict.it Tel. +39 095 7382375; Fax. +39 095 7382397 ABSTRACT Video transmission and P2P networks are causing a tremendous revolution in the current Internet. For this reason performance analysis frameworks are in great demand by Internet Service Providers (ISP) and telecommunications companies in order to re-design their systems so as to support the new paradigms. In this context the target of this paper is to propose an analytical model of a real-time video transmission system using the P2P paradigm to achieve multipoint communication on the current Internet, which is, except for a very limited isles, a unicast network. The system considered is defined in such a way as to maximize the number of users, while guaranteeing a given quality of service level in terms of the maximum number of hops connecting them with the video source. A case study is used to show how the model can be applied to evaluate the performance of the P2P network in a real case, and to support network designers in choosing some network parameters. Keywords: P2P, video transmission, real-time, Markov models, performance evaluation 1. Introduction In recent years, with the widespread deployment of broadband access, multimedia communication has gained an enormous importance. For this reason, in the next generation Internet, interest in multimedia services and video communications in particular is expected to grow significantly. However, most of the multimedia applications which today are being shown great interest by both Internet service providers and users, for example Video on Demand (VoD) and Live Video transmission, are based on distributing multimedia content to a number of receivers simultaneously. In the first generation of such systems the most widely used approach is unicast streaming. According to this approach, multimedia data are transmitted by the source to each client in a point- to-point manner: whenever a new client starts a multimedia session, a dedicated stream is allocated to serve the user till the end of the viewing session. However, as unicast streaming requires a separate streaming bandwidth for each client, the multimedia source and network bandwidth resources consumed will inevitably grow linearly with the user population. This approach is therefore not scalable with the number of users, and is unfeasible even for small groups of users if for example the source accesses the Internet with a low- cost access like ADSL.