High-level simulation of the timely behavior of intermittent systems

Citation
11th International Workshop on Energy Harvesting & Energy-Neutral Sensing Systems (ENSsys 2023)
Contribution to Conference
11th International Workshop on Energy Harvesting & Energy-Neutral Sensing Systems, ENSsys 2023  
Publisher DOI
10.1145/3628353.3628539
Scopus ID
2-s2.0-85180129889
Publisher
Association for Computing Machinery New York, NY
ISBN
979-8-4007-0438-3
Batteryless sensors accelerate the Internet of Things by making wireless sensor applications sustainable. However, their inherent and frequent power failures render batteryless devices less dependable than their battery-driven equivalents. To regain dependability for intermittent systems, energy-aware scheduling strategies are required. Evaluating and comparing such algorithms only on hardware can be tedious and time consuming. Furthermore, environmental impacts and hardware variation affect repeatability and reproducibility. Therefore, we present a high-level modeling approach based on energy flows that allows the simulation of the timely behavior of intermittent systems. In specific, we model a solar-powered batteryless sensor node and use a recorded out-door solar trace as harvest input. To show the usefulness of the approach, we compare the two different operating strategies soft-Intermittent Computing and hard-Intermittent Computing. To validate the model a harvesting power estimation algorithm is implemented on the hardware and compared with the simulation. The results show that the simulation reproduces the intermittent behavior in an appropriate way. Moreover, we find that soft-IC can be a suitable substitute for hard-IC with the benefit of simplified timekeeping during times of power failure.
Subjects
Intermittent Computing,
Simulation of Batteryless Devices
Energy Harvesting
DDC Class
004: Computer Sciences