ReaDmE: Read-Rate Based Dynamic Execution Scheduling for Intermittent RF-Powered Devices

TL;DR

ReaDmE is a dynamic scheduling method that uses RFID read-rate measurements to prevent brown-out failures in intermittently powered devices, enabling more reliable long-term task execution without additional hardware.

Contribution

It introduces a novel read-rate based approach for estimating power channel conditions and a scheduling scheme to improve execution success in CRFID devices.

Findings

20% increase in long-run execution success rate

up to 23% reduction in time overhead

effective without extra hardware or device burden

Abstract

This paper presents a method for remotely and dynamically determining the execution schedule of long-running tasks on intermittently powered devices such as computational RFID. Our objective is to prevent brown-out events caused by sudden power-loss due to the intermittent nature of the powering channel. We formulate, validate and demonstrate that the read-rate measured from an RFID reader (number of successful interrogations per second) can provide an adequate means of estimating the powering channel condition for passively powered CRFID devices. This method is attractive because it can be implemented without imposing an added burden on the device or requiring additional hardware. We further propose ReaDmE, a dynamic execution scheduling scheme to mitigate brownout events to support long-run execution of complex tasks, such as cryptographic algorithms, on CRFID. Experimental results demonstrate that the ReaDmE method can improve CRFID's long-run execution success rate by 20% at the critical operational range or reduce time overhead by up to 23% compared to previous execution scheduling methods.