DOI: https://doi.org/10.36347/sjet.2025.v13i11.003

Pages: 870-877

The extensive implementation of the multi-party analytics cloud computing faces significant security issues especially with sensitive information. The paper presents a confidential computing infrastructure designed to secure multi-party analytics on the Google Cloud Platform, and based on AMD technology SEV-SNP. We have a performance overhead of 22.3 per cent compared to non-homomorphic methods, which is significantly less than the overhead incurred using homomorphic encryption (1,500 per cent) and insecure multi-party computation (800 per cent), and ensures complete data confidentiality throughout computation. The system is able to attain an encryption throughput of 260MB/s, the attestation success rate of 99.98 percent and it can be effectively scaled to support ten subjects. Practical use in experimental assessments of the framework, it is shown to be applicable to privacy preserving analytics through the use of healthcare, financial, and research data and meet regulatory compliance standards.