The simplicity of creating, distributing and communicating content causing an enormous load on the networks. This requires the development of more cost-effective communication. Peer-to-peer (P2P) systems provide a very popular and efficient way of content distribution. The P2P approach becomes increasingly interesting also for video streaming applications. In recent years theoretical and practical solutions have arised that allow to share multimedia content for millions of users at the same time. The performance of mobile devices allows them to participate in streaming and multimedia content distribution applications. Mobile devices typically have limited storage and limited computing resources. There is a new challenge to develop methods that allow the seamless transfer of video and content distribution in mobile environments. Multimedia real-time content distribution is a very complex task, which represents new challenges. We are focusing on the following aspects.
We have developed an open source Java library for deterministic network coding , which also supports other network coding methods, e.g. random linear network coding. It supports P2P content distribution, particularly in mobile environments. The implemented functions suited for devices with limited memory and computing resources, which is typical for mobile devices.
In tracker-based P2P networks, the effectiveness of downloading largely depends on the neighbors of the client. We have developed a new neighbor selection strategy for building the overlay network. Applying randomized downloading strategies in the resulting network faster content distribution can be guaranteed than the previous best upper bound. The speed of the content distribution meets the lower bound up to a constant factor.
We examine, how network coding can be used to improve the transmission in P2P systems. We are developing a P2P video streaming simulator and a prototype system.
Distributed storage systems use redundancy to provide reliable access to data spread over individually unreliable nodes. In case of a node failure a new node must substitute the failing one. We are examining, how network coding can be used to to reduce the repair traffic in distributed storage systems and developing a network coding based distributed storage simulator.
Smart Ubiquitous Content (SmartUC)
Students: