The benefits of peer-to-peer (P2P) systems as vehicles for information dissemination and virtual collaboration are overwhelming. They beat the conventional client-server (C-S) approach in terms of scalability, availability, and network efficiency, to mention a few benefits (see slide 2). Nevertheless, from the management point of view, existing P2P systems are not commercially viable. Key problems are: Operator bypass and legality. P2P services can be easily offered by third-party providers beyond the network operator's ability to control them (simple examples are Skype [1] (offering P2P communication services) and BitTorrent [2] (offering P2P file sharing services). Operator bypass poses problems in terms of management of resources. It is very hard (if not impossible) to monitor and control the traffic incurred by a P2P system, which leads to resource saturation and poor performance (in fact many network providers and campus networks rate limit or even block P2P traffic). The other problem is that the openness and lack of control of existing P2P systems leads more and more often to copyright infringement and lawsuits [3] - i.e., current P2P systems are not suited to distributing copyrighted digital material. Business models and charging. Charging schemes conventionally used in the Telco domain are based on well-known revenue sharing models, interconnection principles and charging mechanisms. These cannot be easily ported onto current P2P systems which bypass the operator (bullet 1) and are self-managed. An effective model based on incentives is adopted in our work. Interoperability. Current P2P systems are mainly proprietary and do not interoperate with each other. Initiatives for standardizing P2P platforms have so far not lead to any definite solution. Lack of interoperability defies the very purpose of P2P - i.e. having a system where any terminal can share resources - and considerably limits their ubiquity. Our P2P-IMS prototype indicates a possible way forward, since the IMS is already the service provisioning platform of choice. Mobile friendliness. Current P2P systems are heavyweight and not suited to thin mobile terminals. This is a considerable limitation given the commercial trends and developments in the area of mobile networks. Our prototype works on mobile UMTS terminals, bringing the power of P2P to the mobile user. Working with a major network operator, we have developed a new approach to managed, operator-mediated P2P service provisioning. Our prototype is built as a new service component of the IP Multimedia Subsystem (IMS) [5], the standardized service provisioning platform of choice of the operators. Here we illustrate the benefits of this approach which brings the advantages of existing P2P systems into the realms of the operator, the service provider, and the digital media production and distribution companies. Our prototype can also be seen as a way to demonstrate how the IMS can be P2P-enabled. The current IMS is based on a C-S architecture and is therefore it not geared for the deployment of large-scale P2P services.
展开▼
