Projects per year
Timely delivery of sensor data is crucial for a wide array of Internet of Things (IoT) applications. Due to the large space- and time-correlation of sensor data, there is a high potential for compression. However, conventional wisdom dictates that compression is at odds with information freshness and timely delivery of data. The reason is that sufficient data needs to be accumulated in order to achieve reasonable compression rates, which introduces additional delays on data transmission. This paper studies a novel approach to perform online compression of data across multiple data sources which achieves significantly better performance in both Age of Information (AoI) and compression for sensor applications. More specifically, we show that our approach can remove the trade-off between these two metrics, particularly, when considering an instantly decodable variant of our approach. We also propose and study techniques to further improve both these metrics by using preset and dynamically created dictionaries at the source nodes. Using real-world data sets, we show that our solution reduces the age of information (by up to a factor of 2.3) and compression ratio (by up to an order of magnitude) with respect to DEFLATE and LZW. Finally, we show that using multiple sources benefits results in an improvement of AoI and compression for each involved source compared to compressing individually.
- Age of Information (AoI)
- generalized data deduplication (DD)
- lossless data transmission reduction
- multiple source nodes
- wireless sensors