Peer-to-peer to drive video multicasting
Innovations in P2P communication have laid the groundwork for whole new ways of using the internet as a service. But the rapid evolution of P2P networks presents a problem for the so-called 'architects of the internet' - the people who develop the hardware, software, middleware and other bits and pieces that make it work. How can quality of service (QoS) be maintained as more and more bandwidth is taken up by data-hungry video applications and services? Can P2P middleware which mediates the networks offer an innovative business model for the future internet? But then where will the telecoms operators fit in? You can look at multicast service delivery as the natural evolution of communications which started as a one-to-one exercise (you talking to your neighbour over the fence or your sister over the phone). But thanks to technological advances, the opportunity of one-to-many communication (through the likes of radio and TV) is now evolving into a many-to-many paradigm through the internet. This is where peer-to-peer networks become important. It is also where the lines between traditional communications providers, operators, broadcast services, etc. get blurred. Innovative business models are needed to embrace these changes, new middleware is needed to ensure QoS, and guidelines are needed to help the various actors settle on a workable and profitable future particularly for P2P-mediated video multicasting, which tends to hog bandwidth. The European P2P-Provideo project set out to provide technical guidelines that will implement P2P multicast transport services in an economical and profitable way, as well as scientific conclusions about the technical choices that warrant scalability and stability for P2P multicast applications. On the technical side, the team has made good progress particularly on the video encoding work, reaching fine granularity scalability (FGS) for video transmission using an MPEG-4 FGS encoder developed by the project which can encode and decode video in real-time. The encoder has been implemented in Visual C++, an IT development environment, using Intel's integrated performance primitives (Intel IPP), software optimised for multimedia communications. 'Using the FGS as video encoding, a multipoint broadcast video transmission framework over a heterogeneous content distribution P2P network has been proposed,' note the project team. It is a live video broadcast platform where a video source distributes a video stream to a number of clients in a multipoint fashion. Multipoint communication has been achieved by applying a P2P approach, configuring a tree-structured overlay network (built on top of another network) where the root is the video source, while the other clients are internal nodes or so-called 'leaves', the researchers explain. The project, which is due to end late 2011, has also developed a revenue model for their system putting users divided into two classes according to how they will pay for the service: with bandwidth (cheap-tariff peers), or with money (full-tariff peers). The model should help service providers design their networks to maximise their revenue while satisfying user requirements for top-quality video streaming. 'On one hand we expect to offer an effective platform for the management of user-provider relationships and for price negotiation of QoS broadband services,' note the project team. 'On the other hand, we will identify the most flexible architecture in terms of adaptive encoding techniques and adaptation logic satisfying the requirements of multicast video services.'
