Approximate statistic query via sampling for IoT data trading and sharing

Fan Yang, Di Zhang, Junqing Le, Jiali Wu

Abstract: With the exponential growth of data, especially in the Internet of Things (IoT) era, paid data-sharing (data trading) has gained attention. Typically, data trading is performed by a third party that purchases and processes data and then sells the aggregated data to data users. However, data trading faces numerous challenges, especially in IoT data trading, where privacy leaks and low data quality have become increasingly prevalent. In this paper, to solve these problems, we introduce ASQ-S, a model designed to support approximate statistic query services of IoT data while ensuring privacy protection, cost-effective, and flexibility. Taking into account diverse backgrounds data broker, we design two sampling methods, BBSM and BFSM. BBSM allows data brokers to reduce service costs when they have prior knowledge of the data being traded, while BFSM provides a cost-effective option when data knowledge is limited. Through extensive experiments, we demonstrate the effectiveness of ASQ-S in various real-world data trading scenarios.