Enhancing Data Privacy and Security in Heterogeneous Edge-to-Cloud Frameworks: Utilizing Confidential Computing for Secure Distributed Processing in Decentralized Networks

Ensuring data privacy and secure processing in heterogeneous edge-to-cloud ecosystems is critical due to the growing prevalence of decentralized data flows and multi-tenant infrastructures. Confidential computing, utilizing Trusted Execution Environments (TEEs), has emerged as a viable solution by enabling computation on encrypted or sensitive data within hardware-isolated enclaves. This paper introduces a Confidential Computing-Enabled Secure Framework (CCESF) that integrates lightweight cryptographic schemes and adaptive trust policies within TEE-based edge-cloud systems. The framework is evaluated using domain-specific datasets aligned with edge-cloud security scenarios [10], [11], rather than outdated intrusion benchmarks. We explicitly detail architectural choices, including a Multi-Activation Function (MAF) variant of LSTM, and assess comparative baselines (e.g., HE, SMPC, DP) through re-implementation with consistent configurations. Furthermore, we conduct cryptographic evaluations and system-level assessments including enclave leakage and side-channel robustness. Empirical results show classification accuracies of 99.02% (IoT -edge dataset) and 98.88% (cloud workload dataset), validated with stratified 10-fold cross-validation and reported with confidence intervals. Our findings highlight the practical viability of CCESF for secure and scalable edge-cloud computing.