Privacy-preserving and fault-tolerant aggregation of time-series data without TA

The increase in popularity and usage of the Internet of Things (IoT) applications, along with big data, has highlighted time-series data aggregation. Data is continuously and periodically generated in a time-series scenario and then transported to the aggregator for analysis. Data aggregation is a helpful operation to preprocess data, where a group of users sense the time-series data. However, some security and privacy issues still need to be solved. Many traditional privacy-preserving solutions cannot support fault tolerance, a vital feature in time-series scenarios. Moreover, a trusted authority is difficult to build in the real world. This paper proposes a privacy-preserving time-series data aggregation scheme without TA. The proposed scheme can also compute arbitrary aggregate functions and achieve fault tolerance for enhancing data aggregation's reliability and scalability. Security analysis demonstrates that our proposed scheme achieves forward secrecy and fault tolerance. We also conduct thorough experiments based on a simulated data aggregation scenario to show the scheme's computation and communication efficiency.