Coded Computing and Cooperative Transmission for Wireless Distributed Matrix Multiplication

Consider a multi-cell mobile edge computing network, in which each user wishes to compute the product of user-generated data matrix with network-stored matrix. This is done through task offloading by means input uploading, distributed at nodes (ENs), and output downloading. Task may suffer long delay since servers some ENs be straggling due random computation time, wireless channels experience severe fading interference. paper aims investigate interplay among upload, computation, download latencies during process high signal-to-noise ratio regime from an information-theoretic perspective. A policy based on cascaded coded coordinated cooperative interference management uplink downlink proposed proved approximately optimal for sufficiently large upload time. By investing more time transmission, creates redundancy ENs, can reduce enabling use computing, as well via transmitter cooperation. Moreover, allows traded Numerical examples demonstrate that improve over existing schemes significantly reducing end-to-end execution