Network coding for Rapid and Reliable Data Delivery (R2D2)

11^th March 2015, 11:00 Ashton Lecture Theater
Dr Ioannis Chatzigeorgiou
School of Computing and Communications
Lancaster University

Abstract

Network coding has the potential to significantly improve network reliability by mixing packets at a source node or at intermediate network nodes prior to transmission. In this talk, the concept of network coding will be presented and extended to the case of data comprising multiple layers of different importance levels and require unequal error protection. The idea of unequal error protection will then form part of a resource allocation framework, whose objective can be either provider-centric or user-centric. In the former case, the provider optimises the number of transmitted coded packets and the adopted modulation and coding scheme in order to offer a service to a minimum fraction of users without violating an existing service lever agreement. In the latter case, the aim is to maximise the ratio between the number of recoverable layers by the users (users' profit) and the total number of coded packet transmissions (provider's cost). The impact of the adopted network coding method (i.e., systematic or non-systematic transmission) on performance and the effect of sparse network coding on packet transmissions and decoding complexity will also be briefly discussed. The talk will conclude with a mapping of the aforementioned practical problems to theoretical problems related to the determination of the rank of constrained random matrices over Galois fields.