Optimal Resource Allocation in Joint Secret Key Generation and Data Transfer Schemes

Due to computational complexity and latency constraints in the nodes of many IoT systems, alternatives to public key encryption are sought for session key generation schemes. In this work we investigate novel cross-layer security protocols in which session keys are generated at the physical layer using standard techniques of secret key generation (SKG) from shared randomness. In this framework, we study the optimal power allocation in block-fading additive white Gaussian noise (BF-AWGN) channels, with short-term power constraints, when a subset of the subcarriers are used for SKG and the rest for data transmission. Fixing the amount of data that can be transmitted with a single key, allows us to first identify the optimal subset of subcarriers that should be devoted to SKG and the respective power allocation policy which, depending on the available overall power, might not be unique. Subsequently, a further step is taken in our analysis to account for the impact of the proposed power allocation with long-term power-constraints.