Privacy-preserving and verifiable aggregation for multi-task and multi-dimensional data

Abstract With the widespread adoption of smart devices, vast amounts of multidimensional data are continuously generated across various sectors. Integrating and evaluating this data can provide strong support for decision-making. However, existing aggregation schemes face several challenges when handling multidimensional data tasks from multiple requesters, including single points of failure, privacy leakage, lack of result validation, and inefficiency. This paper mitigates the challenges by integrating packed secret sharing, Kate, Zaverucha and Goldberg (KZG) commitments and the Chinese Remainder Theorem to design a data aggregation scheme. For the first time, we achieve both multi-task and multi-dimensional data aggregation within a single data transmission, ensuring data integrity and privacy, while keeping the overhead minimal. Users can provide data for only certain dimensions, thereby safeguarding personal data privacy while emphasizing the user-centric approach to data control. By incorporating multiple fog nodes (FN) and blockchain, the data aggregation scheme is implemented in a decentralized manner. Both data users and FNs are accountable, ensuring the data authenticity and the correctness of the aggregation process. Moreover, it is fault-tolerant, making the system robust. Rigorous security analysis and extensive experimental evaluations demonstrate that the proposed approach excels in computational efficiency.