Skip to Main Content
We consider i) the overhead minimization of maximum-distance separable (MDS) storage codes for the repair of a single failed node and ii) the total secure degrees-of-freedom (S-DoF) maximization in a multiple-access compound wiretap channel. We show that the two problems are connected. Specifically, the overhead minimization for a single node failure of an optimal MDS code, i.e. one that can achieve the information theoretic overhead minimum, is equivalent to maximizing the S-DoF in a multiple-access compound wiretap channel. Additionally, we show that maximizing the S-DoF in a multiple-access compound wiretap channel is equivalent to minimizing the overhead of an MDS code for the repair of a departed node. An optimal MDS code maps to a full S-DoF channel and a full S-DoF channel maps to an MDS code with minimum repair overhead for one failed node. We also state a general framework for code-to-channel and channel-to-code mappings and performance bounds between the two settings. The underlying theme for all connections presented is interference alignment (IA). The connections between the two problems become apparent when we restate IA as an optimization problem. Specifically, we formulate the overhead minimization and the S-DoF maximization as rank constrained, sum-rank and max-rank minimization problems respectively. The derived connections allow us to map repair strategies of recently discovered repair codes to beamforming matrices and characterize the maximum S-DoF for the single antenna multiple-access compound wiretap channel.
Date of Conference: Sept. 29 2010-Oct. 1 2010