Network Coding Based Multimedia Streaming in Content Centric Networks

Information Centric Networking architectures (ICN) have recently gained significant attention in the research community, as they promise to revolutionize the way data is exchanged in the Internet. They move from the traditional paradigm of Internet communication using IP addresses towards using names as addresses. This is motivated by the fact that when users browse the Internet, they care only about the data content and not where the content is stored. On the contrary, the IP model of communication focuses on where the data is located. Several problems are associated with the current IP network architecture like usability, performance, security and resilience to mobility. To cope with some of these limitations, content distribution networks (CDN) and peer-to-peer architectures have been proposed. These methods mainly deal with the scalability issue and attempt to exploit better the available network resources. CDN and P2P could be seen as a first step towards ICN.

Network coding has been presented a decade ago as an efficient technique for heterogeneous both wired and wireless overlay networks to increase the throughput, decrease the delay, enhance resilience, remove the need for coordination between the network nodes etc. There are two major classes of network coding algorithms namely Linear Network Coding (LNC ) and Random Linear Network Coding (RLNC). Both methods operate in finite fields. LNC decides about the coding operations centrally, although there are some decentralized designs, whereas RLNC randomly performs operations in finite fields and has only a small performance penalty compared to LNC when operations are in large finite fields. Network coding is interesting for multimedia communication. The challenge with multimedia is that data is often scalable and data delivery should respect the tight decoding deadlines.

In this project, we envisage the design of novel network coding methods that will promote the use of ICN. We will build our techniques on the Content Centric Networking (CCNx) implementation, since it has many advantages like hierarchical prefixes and is open source. Some abstract ideas regarding the use of network coding in CCN have been very recently discussed. It mainly provides some examples motivating the appropriateness of network coding for the ICN framework, rather than specific solutions. In our perspective, specific problems should be resolved prior to employing such technologies. Specifically, open challenges are: what kind of prefixes should be used, security issues, where to cache information, how one can deal with multiple concurrent sessions accessing the network, could data correlation be exploited? The target of our project is twofold: (a) design network coding techniques that will improve the perceived quality of services and (b) propose an architecture for CCN appropriate for network coding enabled systems. We will focus on multimedia streaming applications, as it is the main source of data traffic in today’s Internet. We will further target on the employment of our methods in social networks deployed when users want to share multimedia data. We believe that the ICN paradigm fits well into the framework of multimedia communication over social networks as users can take advantage of multiple interfaces to acquire the multimedia data faster and exploit efficiently the cached data as typically many users seek for the same multimedia data. We are convinced that the employment of network coding in CCN will accelerate the data delivery, improve multimedia quality, enable better the available resources, and revolutionize the caching strategies in CCN framework by considering data importance.