A Dual-Threshold Scheme Along with Security Reinforcement for Energy Efficient Nonvolatile Processors

Abstract

With the increasing scale and decreasing size of the Internet of Things (IoTs) devices, energy harvesting systems have been proposed to power the systems instead of batteries. Addressing the problem that harvested energy is unstable, nonvolatile processors (NVPs) have been proposed to hold intermediate data and avoid frequent program restarting from the beginning. However, NVPs often suffer frequent backup and recovery operations, wasting a lot of energy and system resources. To further improve the performance of NVPs, this paper proposes a dual-threshold method to maximize execution progress by enabling a system to hibernate to wait for power resumption instead of backing up data directly upon power interruptions. In particular, the appropriate retention and backup thresholds are discussed in details in order to achieve the goal of minimizing power failures and maximizing computation progress. In the meantime, the possible attacks to NVPs with dual-threshold and solutions combating these threats are discussed to guarantee NVP's security. The evaluation results show an average of up to 82.3% reduction on power failures and 1.5x speedup on forward progress by the proposed dual-threshold method compared to the conventional single threshold scheme.