Purpose: The purpose of this study is to examine the application of Federated Machine Learning (FML) within hybrid cloud environments, where public and private infrastructures coexist. While FML inherently enhances privacy by keeping data localized, its deployment in hybrid clouds introduces complex challenges regarding data security, compliance, and trust. This article aims to identify the critical trade-offs between model accuracy, computational efficiency, and privacy preservation, while proposing a framework to address these issues. Methodology: Evaluating adversarial attacks and data leakage risks specific to distributed and hybrid cloud contexts. Investigating privacy-preserving techniques such as differential privacy, secure multiparty computation, and trusted execution environments, with a focus on their scalability and performance in hybrid deployments. Designing a security-aware framework that balances trust management, policy enforcement, and data protection across hybrid cloud infrastructures. Conducting scenario-based analyses to demonstrate how organizations can implement federated learning within hybrid clouds while meeting compliance and data sovereignty requirements. Findings: The findings reveal that federated learning in hybrid clouds can provide significant benefits in terms of privacy and regulatory compliance. Organizations must balance performance metrics, such as model accuracy and training efficiency, with stringent security requirements. Differential privacy and secure multiparty computation offer strong protection but may degrade efficiency, while trusted execution environments present a middle ground with practical benefits for hybrid scenarios. Unique Contribution to Theory, Policy and Practice: The proposed security-aware framework supports adaptable and resilient implementations, helping organizations enforce policies, manage trust relationships, and safeguard sensitive data. Effective adoption requires aligning technical safeguards with regulatory mandates, ensuring that privacy-preserving strategies remain adaptable across evolving multi-cloud ecosystems.
Sri Ramya Deevi (Tue,) studied this question.