In recent years, multi-tenant cloud computing has become the default architecture for modern data services, yet the same shared infrastructure that makes it cost-effective also creates a persistent privacy risk, due to tenant data is typically encrypted only at rest and in transit, leaving it exposed in plaintext the moment it is processed. This gap creates opportunities for unauthorized access through misconfigured permissions, compromised hypervisors, or malicious co-tenants. This study addresses this gap by designing and implementing a cryptographic data isolation model using Homomorphic Encryption (HE) to protect tenant data throughout its entire lifecycle, including during computation. The model’s design is grounded in the Brakerski-Fan-Vercauteren (BFV) scheme for lattice-based, quantum-resistant key management, addition, and multiplication over ciphertext, and was implemented and evaluated using the Cheon-Kim-Kim-Song (CKKS) homomorphic scheme within a Python application layer backed by a MySQL database Server. Furthermore, homomorphic key pairs were generated and used to encrypt and decrypt both user profile and financial transaction data, with encryption and decryption times measured across ciphertext sizes ranging from 4,800 to 50,000 bytes. Experimental results show that encrypted values remained completely unreadable regardless of message length, confirming strong confidentiality, while encryption and decryption times scaled predictably with data size, from 0.35ms/0.12ms at 4,800 bytes to 3.6ms/1.9ms at 50,000 bytes. A comparative analysis against Rivest-Shamir-Adleman (RSA), Data Encryption Standard (DES), Advanced Encryption Standard (AES), and Elliptic Curve Cryptography (ECC) further showed that although HE carries a heavier computational cost, as it uniquely allows computation directly on encrypted data, making it a suitable foundation for privacy-preserving multi-tenant cloud systems.

This study addressed one of the most significant security challenges in multi-tenant cloud computing environment by developing a privacy-preserving cryptographic model capable of protecting tenant data throughout its entire lifecycle. Although conventional encryption techniques provide adequate protection for data at rest and during transmission, they require data to be decrypted before processing, thereby exposing sensitive information within the cloud environment. To overcome this limitation, this study integrated a homomorphic encryption model into the cloud data processing workflow, enabling computations to be performed directly on encrypted data without revealing the underlying plaintext. By eliminating the need for decryption during processing, the proposed model strengthens tenant isolation and significantly enhances data confidentiality within shared cloud infrastructures. If there is no conflict of interest, authors should state that “The author(s) declare(s) that there is no conflict of interest regarding the publication of this paper. The proposed model was implemented using the BFV and CKKS homomorphic encryption scheme in a Python environment with MySQL serving as the database management system. Experimental evaluation confirmed the successful generation of homomorphic public and private key pairs, while user identity information and financial transaction records were securely transformed into ciphertext that remained unintelligible without the corresponding decryption key. The decryption process consistently reconstructed the original plaintext without any loss of data integrity, demonstrating the correctness, reliability, and effectiveness of the proposed cryptographic framework. The experimental results further demonstrated that encryption and decryption times increased gradually and predictably as the size of the input data increased. Even for larger message sizes, processing times remained within a few milliseconds, indicating that the computational overhead introduced by HE is manageable for realistically sized cloud data records. Comparative evaluation with the Rivest-Shamir-Adleman (RSA), Data Encryption Standard (DES), Advanced Encryption Standard (AES), and Elliptic Curve Cryptography (ECC) further highlighted the strengths and limitations of our proposed approach. Although HE incurred greater computational cost than these conventional encryption algorithms, it remained the only technique capable of performing computations directly on encrypted data while preserving data confidentiality throughout the computation process. This capability represents a significant advantage for multi-tenant cloud environments where preventing unauthorized access to tenant information during processing is a fundamental security requirement. The findings of this study demonstrate that HE provides a practical and effective foundation for privacy-preserving cloud computing. Rather than depending solely on access control mechanisms or organizational security policies, the proposed model embeds privacy protection directly within the cryptographic process itself. This approach substantially reduces the opportunities for data exposure arising from insider threats, system misconfigurations, unauthorized access, or malicious attacks, thereby providing stronger guarantees for tenant data confidentiality in shared computing environments. Although the proposed model achieved its intended objectives, the evaluation was conducted under controlled experimental conditions using individual records and did not consider large-scale concurrent processing involving multiple tenants. Consequently, the performance of the model under production-scale cloud workloads remains an area for further investigation. Future studies should evaluate the scalability of the proposed model in large multi-tenant cloud environments, investigate hybrid cryptographic approaches that combine homomorphic encryption with efficient symmetric encryption techniques to reduce computational overhead, and assess the model against a broader range of sophisticated cyber threats, including side-channel and inference attacks. Such investigations would provide additional evidence of the model’s suitability for deployment in real-world cloud computing environments. This study has demonstrated that the computational overhead associated with Homomorphic Encryption (HE) is outweighed by its unique ability to preserve data confidentiality during computation. The proposed model therefore represents a significant advancement toward secure and privacy-preserving multi-tenant cloud computing by providing cryptographic protection that extends beyond storage and transmission to include active data processing. Consequently, the study offers a practical and scalable framework for strengthening tenant isolation and improving trust in cloud computing environments where data privacy remains a critical concern.

[1] M. Kansara, “Advancements in cloud database migration: Current innovations and future prospects for scalable and secure transitions,” Sage Science Review of Applied Machine Learning, vol. 7, pp. 127-143, 2024. [2] W. Hashim and N. A.-H. K. Hussein, “Securing cloud computing environments: An analysis of multi-tenancy vulnerabilities and countermeasures,” Shifra, vol. 2024, pp. 8-16, 2024. [3] B. A. Sekti, “Data Privacy and Cybersecurity in Wind Power Systems,” in AI-Powered Analysis, Modeling, and Monitoring of Wind Energy Systems, ed: IGI Global Scientific Publishing, 2026, pp. 335-366. [4] N. S. Gaddapuri, Cloud Computing in Regulated Financial and Healthcare Systems (Architecture Patterns, Security, and Compliance Frameworks): Geh press, 2026. [5] B. R. Louassef, N. Chikouche, H. Mrabet, and S. Belguith, “A privacy-preserving scheme for iot healthcare systems using blockchain and lattice-based cryptography,” The Journal of Supercomputing, vol. 82, p. 295, 2026. [6] O. E. Taylor and I. N. Davies, “A Model for Enhancing Security and Privacy in Pervasive Computing using Homomorphic Encryption ” International Journal of Computer Sciences and Engineering, vol. 13, pp. 21-29, 2025. [7] B. O. Eboseremen, A. O. Ogedengbe, E. Obuse, O. Oladimeji, J. O. Ajayi, A. O. Akindemowo, et al., “Secure data integration in multi-tenant cloud environments: Architecture for financial services providers,” Journal of Frontiers in Multidisciplinary Research, vol. 3, pp. 579-592, 2022. [8] S. K. Henge, R. Rajakumar, P. Prasanna, A. Parivazhagan, Y.-C. Hu, and W.-L. Chen, “Multi-layered access control based auto tuning relational key implications in enterprise-level multi-tenancy,” Multimedia Tools and Applications, vol. 84, pp. 17805-17836, 2025. [9] P. Kumar and A. K. Bhatt, “HE-AO: An Optimization-Based Encryption Approach for Data Delivery Model in A Multi-Tenant Environment,” Wireless Personal Communications, vol. 138, pp. 1329-1350, 2024. [10] P. Mehta, “SECURING CLOUD ENVIRONMENTS WITH HOMOMORPHIC ENCRYPTION,” International Research Journal of Advanced Engineering and Technology, vol. 1, pp. 25-30, 2024. [11] S. I. Serengil and A. Ozpinar, “Lightphe: Integrating partially homomorphic encryption into python with extensive cloud environment evaluations,” arXiv preprint arXiv:2408.05219, 2024. [12] J. Wu, T. Sun, F. Luo, H. Wang, and W. Zhang, “Secure multi-key homomorphic encryption with application to privacy-preserving federated learning,” arXiv preprint arXiv:2506.20101, 2025. [13] Y. H. Chan, H. Yang, S. Shen, X. Fan, S. Lyu, P. S. Hung, et al., “HHEML: Hybrid Homomorphic Encryption for Privacy-Preserving Machine Learning on Edge,” arXiv preprint arXiv:2510.20243, 2025. [14] M. A. Junior, P. Appiahene, O. Appiah, and K. Adu, “Cloud data privacy protection with homomorphic algorithm: a systematic literature review,” Journal of Cloud Computing, pp. 1-26, 2025. [15] J. Wang and Y. Wang, “Privacy Protection Optimization Method for Cloud Platforms Based on Federated Learning and Homomorphic Encryption,” Sensors, vol. 26, p. 890, 2026. [16] D. S. Ene, I. N. Davies, G. F. Lenu, and I. B. Cookey, “Implementing ECC on Data Link Layer of the OSI Reference Model,” SSRG International Journal of Computer Science and Engineering, vol. 8, pp. 12-16, 2021.