DR-MDS: An Energy-Efficient Coding Scheme in D2D Distributed Storage Network for the Internet of Things

Abstract

The Internet of Things (IoT) is a promising technology for the collection data among the seamless connectivity with massive physical devices for future 5G networks. While device-to-device (D2D) communication is an emerging technique, which potentially tends to promote the machine-type communications developments of the IoT. In this paper, we consider a wireless distributed storage network where the IoT devices cache decentralized data in each memory and use D2D links to share the whole content, and to save the bandwidth and energy consumptions compared to traditional base station cellular network. We analyze and compare the energy costs of regenerating codes and maximum distance separable (MDS) codes for data downloading and repairing processes with the different parameter configuration. We also define a target named D2D link availability to measure the download and repair efficiency, and obtain the optimal repair intervals of various schemes for the lowest energy cost. Furthermore, we propose a novel caching scheme, termed double replication MDS (DR-MDS), which can modify the repair bandwidth by utilizing more storage occupation, and improve the D2D link availability in the system. The simulation results have shown that the DR-MDS caching significantly outperformed the MDS caching in the energy saving, that also verifies the practicability and the feasibility of our D2D distributed caching system for the low-power IoT scenarios.