Peer-to-Peer networks have already proven themselves indispensable for file sharing applications, however the challenge of successful streaming over P2P still remains unsolved.
The goal of this research project is to build a test-bed for the streaming of scalable video over P2P networks and to investigate how the adaptability of scalable video codecs can be employed in P2P environments.
There are two big advantages of P2P streaming in general:
- Huge bandwidth cost reductions for content providers
- More social and more democratic delivery of user-created content
Currently, almost all of the video streaming over the Internet is being done in client-server fashion over websites such as YouTube for user generated content or content providers' own websites for copyrighted content. Although this architecture has been quite successful so far, as the amount of delivered content keeps increasing and Internet video becomes more mainstream, the server bandwidth costs of providers are bound to explode, especially when a large number of users require the same content. P2P Streaming for those applications can reduce the bandwidth costs by distributing the bandwidth load among all participating peers.
These bandwidth cost reductions do not only apply to big services like YouTube. All users in the network benefit from the same effect. Independent video producers, for instance, can deliver their newest productions to much wider audiences with the same upload bandwidth than it were possible in a server-client architecture, thus reducing the barrier of entry for people who want to make their content available for streaming.
Although, unlike file-sharing applications, P2P streaming applications are less sensitive to packet losses, there are also challenges for a wide-scale deployment of P2P streaming:
Strict play-out deadlines: If a packet is not delivered before its play-out deadline it is useless. Therefore delay and jitter have to be kept check. And packet losses due to peer churn need to be minimized. If there is enough buffer capacity available, a peer can/should query other peers for missing packets.
Asymmetrical up- and downlink bandwidths: Due to popular aDSL connections most peers have enough downlink but limited uplink bandwidth. Therefore they can watch a stream but cannot contribute capacity effectively.
- TU Berlin: Engin Kurutepe (Coordinator)
- QMUL: Naeem Ramzan
- University of Hannover: Zhijie Zhao
- EPFL: Jong-Seok Lee, Francesca De Simone
- Z. Zhao, J. Ostermann, "Video Streaming Using Standard-Compatible Scalable Multiple Description Coding Based on SVC", ICIP 2010
- Z. Zhao, J. Ostermann, H. Chen, “Low Complexity Multiple Description Coding for the Scalable Extension of H.264/AVC”, Picture Coding Symposium (PCS 2009), Chicago IL, USA, May 2009.
- S. Asioli, N. Ramzan, E. Izquierdo, “Efficient Scalable Video Streaming over P2P Network”, Int. ICST Conf. on User Centric Media (UCMedia), Venice, Italy, Dec 2009.
- Z. Zhao, J. Ostermann, “Error Concealment in the Net- work Abstraction Layer for Medium Grain Scalability of SVC”, VCIP 2010, Huang Shan, An Hui, China, July 2010.