Compound Compression Method for Gathering Traffic of IoT/CPS Data
Kazuhito Matsuda, Makoto Kubota · 2019
These days, many kinds of sensors are connected to the Internet. They measure the real world from multiple aspects, and the measured data is used to optimize commercial activity and social infrastructure. This trend is called IoT/CPS, which stands for Internet of Things and Cyber Physical Systems. The data are divided into two types: "heavy data" such as multimedia data and "light data" such as numerical values and strings. Demand exists to gather light data to data centers; however, traffic volume becomes a challenge. In this paper, we propose an efficient compression method targeting light data. The proposed method utilizes the structure of light data, which have "constant," "variable," and "timestamp" fields, to create an efficient dictionary of symbols and perform column-based compression. By evaluating multiple kinds of light data sources, the proposed method can improve the compression ratio by 24% compared with a zip (Deflate) compression and achieve traffic reduction.