Lightweight Compression for Severely Constrained IoT Devices

Conference: European Wireless 2022 - 27th European Wireless Conference
09/19/2022 - 09/21/2022 at Dresden, Germany

Proceedings: European Wireless 2022

Pages: 6Language: englishTyp: PDF

Authors:
Vestergaard, Rasmus; Zhang, Qi; Lucani, Daniel E. (DIGIT and Dept. of Electrical and Computer Engineering, Aarhus University, Denmark)
Techel, Johannes (Dept. of Electrical and Computer Engineering, Bundeswehr University Munich, Germany)

Abstract:
In this paper, we propose S-BP (Sensor Bit-Packing), a lightweight data compression scheme with a tiny memory footprint. Due to its low resource usage, the method is particularly suited for resource-constrained IoT sensor nodes. SBP excels for small data windows in the tens or hundreds of bytes, enabling both compression and energy savings even when collected data must be transmitted frequently. We implemented the proposed algorithm for Contiki OS, proving its suitability for severely constrained devices and allowing us to use realworld data to evaluate it in terms of the achieved compression, processing speed, and energy usage. For sensor data, S-BP achieves comparable and even better compression compared to more resource-intensive compressors designed for powerful devices. We also experience that energy savings are achieved as soon as the compressed size is just 2 % less than the original.