Weakly Secure Coded Distributed Computing
Ruimin Zhao, Jin Wang, Kejie Lu, Jianping Wang, Xiumin Wang, Jingya Zhou, Chunming Cao · 2018
In recent years, edge computing has become a hot topic because it can significantly reduce the latency of network services by utilizing the computation resources at the edge of the network. To further improve the latency performance, several coded distributed computing (CDC) schemes have been proposed, in which computation and communication resources can be used to trade off the delay. To design CDC scheme, data shuffling among computing nodes has been studied but the security issue in data shuffling has not been fully investigated. Therefore, in this paper, we investigate secure data shuffling in edge computing. Specifically, we first formulate a weakly secure edge computing problem, namely, the secure coded distributed computing (SCDC) problem. We then theoretically analyze the existence of weakly secure data shuffling. Based on the analysis, we propose a redundancy reduction (RR) algorithm to assign computing tasks to distributed nodes to minimize the latency with security guarantee, where we prove the correctness of the algorithm and the reduction ratio when RR is compared to existing CDC schemes. Finally, we conduct extensive simulation experiments that confirm the efficiency of the proposed RR scheme.