• Monday, 27 June 2016, 14:00-17:30, Poseidon B

Speakers

• Marcus Greferath (Aalto University School of Science, Finland)
• Ángeles Vazquez-Castro (Autonomous University of Barcelona, Spain)

Motivation and Context

The fundamental motivation of this proposal is the dissemination of the work and results of the COST Action IC1104 with title Random Network Coding and Designs over GF(q). This COST Action is coming to the end in April 2016 after four successful years during which a number of results of very high interdisciplinary nature have been produced and published. The tutorial will provide an account of the deep fundamentals of network coding, coding over network coding and latest results. Practical applications will be also introduced showing how network coding technology can provide actual solutions to technological challenges in the framework of 5G trends in networking. We believe this interdisciplinary tutorial can attract the innovation community at EUCNC, both from industry and academy.

Structure and Content

The tutorial is proposed to be structured as follows.

1. Introduction to network coding and coding for network coding
   The concept of network coding can be dated back to Yeung and Zhang in 1999. The inception of the concept was inspired by low earth satellite (LEO) communication networks and the general problem studied was a multiple-source coding problem. The notion was later established by Rudolf Ahlswede, Ning Cai, Shuo-Yen Robert Li and Raymond W. Yeung in 2000 in a landmark paper where it is shown the potential power of network coding for increasing the amount of information transferred across multicast networks (error-free networks, possibly with link-failures). Thereafter, a large number of results followed with seminal contributions including (optimal) low-complexity distributed random coding algorithms and algebraic models for network coding design. While error-free networks constitute a relevant scenario, the concept of network coding was soon extended for networks with errors and erasures thus including wireless networks. In this case, information needs to be encoded to be transmitted over the network with nodes implementing network coding.
2. Fundamentals of Coding for Network Coding
   In 2008, a seminal and price-winning paper by R. Koetter and F. Kschischang set mathematical foundations for such problem. They proposed two separated problems: the network coding problem for maximizing the network flow (which can be solved assuming error-free links) and the problem of coding for error correction (which can rely on classic coding techniques). Accordingly, their proposed solution was a non coherent transmission model, i.e. neither the source node nor the receiver nodes are assumed to have knowledge of the underlying network coding coefficients. Further, they proposed codes as collections of subspaces (for which distance metric driven constructions are developed). The enormous novelty with respect to classic block codes is that while in classical coding theory codewords are vectors, in subspace codes each codework is itself an entire vector space. This coding has since then attracted much interest among experts in diverse fields, like coding theory, cryptography, and information theory. This particularly stems from the fact that the technology potentially covers contemporary communication and computation environments like the Internet and Cloud Services.
3. From theory to practical design and performance
   The concept of network coding is clearly applicable to real packet networks for the control and optimization of the information flow. Moreover, its applicability is not confined to a single layer in the protocol stack. Experimental results are available proving its applicability at the physical layer (complex coefficients), IP and TCP layers, and of course at the application layer. A wide variety of potential applications can be also drawn within the context of upcoming novel 5G networking trends such as network sofwarization, clodification or virtualization. Even if powerful, such a broad applicability also poses a challenge to a unified design approach over different communication networks and systems. Within the COST Action, a proposal of architectural design for practical design of network coding has been developed. It essentially identifies the several functionalities required for the design of network coding to fulfill different vertical and horizontal design networking objectives ranging e.g from differentiated traffic engineering in the core to energy efficiency at the edge.