Cloud computing has transformed how organizations store and manage their data, with more enterprises moving their operations to cloud-based platforms each year. Yet this shift to cloud storage, despite its operational benefits, introduces serious concerns about protecting sensitive information within these complex systems. The involvement of external service providers who cannot always be fully verified, combined with potential vulnerabilities from third-party access and security breaches, creates multiple pathways for data compromise or unauthorized disclosure.
The research presents an integrated security architecture designed to safeguard data privacy within cloud systems. This architecture combines several protective mechanisms, including techniques for marking data flows, methods to obscure data structures, and controlled access protocols for sensitive information. Testing results demonstrate that the data marking approach outperforms traditional K-anonymity techniques by 18.5% in processing efficiency, while successfully defending against various attack scenarios, including attempts to manipulate data subsets through incremental changes, selective targeting, or modification attacks.
The framework develops an integrity verification system for replicated data that accommodates real-time modifications. This system employs cryptographic signatures combined with randomization techniques to enable both secure validation and flexible data updates. By leveraging ElGamal cryptography alongside homomorphic computing principles, the solution ensures consistency across distributed data replicas housed in various server locations, achieving this protection with minimal processing demands.
A key innovation lies in combining a layered tree-based data architecture with randomization masking methods, resulting in substantial improvements to both verification speed and security levels. The system incorporates HMAC authentication protocols and IO-based obfuscation strategies to create an economical third-party verification mechanism. This approach achieves an 80% reduction in processing requirements when compared to existing solutions.
Through systematic experimental validation, the framework demonstrates effectiveness in balancing data loss rate with privacy protection. When privacy tolerance increases, data availability improves while maintaining security. The multi-tenant cloud computing environment testing shows that data obfuscation mechanisms successfully protect against malicious attacks without significantly increasing storage layer load.
Contributing to this research is Yu Liu, who currently serves as Senior Financial Analyst at 365 Retail Markets in Troy, Michigan, with prior experience as Audit Senior Assistant at Deloitte in Detroit. Liu holds a Master of Accounting (STEM) from the University of Michigan Ross School of Business and a Bachelor’s in Financial Management from Beijing Technology and Business University. Liu’s expertise spans financial data management, automated system design, and implementation of internal controls.
Liu’s professional background in the financial audit industry has directly informed this research. Through extensive experience in evaluating revenue recognition under ASC 606, consolidating intercompany transactions, and eliminating internal control deficiencies, Liu brings practical insights into how audit frameworks must balance security with operational efficiency. This unique combination of technical knowledge in data systems and a deep understanding of audit methodologies has enabled the development of a framework that meets both regulatory compliance requirements and practical business needs.
This work provides organizations with practical tools to protect data privacy in cloud environments while maintaining operational efficiency. The framework addresses critical security challenges in cloud computing adoption and supports the continued digital transformation of enterprises across industries.