A Privacy-Preserving Framework for Big Data Analytics in Edge-Cloud Data Centers

IoT sensors used for real-time monitoring will generate enormous amounts of data, hence new big data analytics (BDA) techniques are needed. All device-generated data is delivered to distributed edge servers, in contrast to the traditional architecture. Technically speaking, this new paradigm offers a robust, flexible, and affordable way of transferring data from mobile devices with constrained energy, storage, and computational capacity to remote edge-cloud data centers. However, in such distributed contexts, security vulnerabilities and deterioration of QoS performance remain evident and open challenges. Three significant contributions are made by this research. First, we propose a lightweight protocol solely comprised of bitwise exclusive operator (XOR) and one-way hash functions for key agreement and mutual authentication. Second, we employ K-anonymity as well as T-closeness for data publication anonymization. Third, for privacy-preserving outsourced computation, we develop a multi-party computation (SMC) protocol using the Paillier cryptosystem. The simulation outcomes demonstrate that, while using distant services, the suggested framework improves data security and lowers the possibility of unauthorized data disclosure.