Private and rateless adaptive coded matrix-vector multiplication

Abstract Edge computing is emerging as a new paradigm to allow processing data near the edge of network, where typically generated and collected. This enables critical computations at in applications such Internet Things (IoT), which an increasing number devices (sensors, cameras, health monitoring devices, etc.) collect that needs be processed through computationally intensive algorithms with stringent reliability, security latency constraints. Our key tool theory coded computation, advocates mixing tasks by employing erasure codes offloading these other for computation. Coded computation recently gaining interest, thanks its higher smaller delay, lower communication costs. In this paper, we develop private rateless adaptive (PRAC) algorithm distributed matrix-vector multiplication taking into account (1) privacy requirements IoT (2) heterogeneous time-varying resources devices. We show PRAC outperforms known secure methods when are heterogeneous. provide theoretical guarantees on performance comparison baselines. Moreover, confirm our results simulations implementations Android-based smartphones.