Comparative study of compression algorithms on time series data for IoT devices

This Master’s thesis evaluates the performance of lightweight compression algorithm aimed for IoT sensor devices. These devices are most often battery driven and produce large amounts of sensor data which is sent wirelessly. Compressing the data could be a tool in decreasing the power usage of these devices. The algorithms were evaluated based on their compression ratio, memory consumption and CPU usage. The results showed that the algorithms DRH, improved RLBE and A-LZSS performed the best according to different criteria. DRH gave the most balanced performance on all metrics, while improved RLBE is aimed towards datasets with steeper changes, and A-LZSS towards datasets with reoccurring sequences