RAMSES: New HPC system delivers security without sacrificing speed
Encryption at rest, in transit, and in use—with only minimal performance loss.
Traditional HPC architectures prioritize raw speed, often at the expense of security—a trade-off that becomes problematic for fields like life sciences, where sensitive data (e.g., genomic or patient records) must be processed at scale. Researchers from multiple European institutions have now introduced RAMSES (Research Accelerator for Modeling and Simulation with Enhanced Security), an HPC system designed from the ground up to deliver high performance within a robust security framework. The system leverages AMD processor-based hardware memory encryption, IBM Storage Scale for file-level encryption, and the Thales CipherTrust manager to ensure data is encrypted at rest, in transit, and even in use—covering the entire data lifecycle.
RAMSES also implements advanced OS hardening, a multi-layered security architecture, and mandatory multi-factor authentication, helping it comply with major data protection standards such as GDPR, ISO/IEC 27001, and FIPS. Benchmark tests in the biomedical sector demonstrate that the performance impact of these security measures is limited, proving that the conflicting requirements of speed and security can be reconciled. The result is a coherent, flexible, and user-friendly system that opens the door to secure large-scale processing of sensitive data without forcing researchers to sacrifice computational performance.
- RAMSES combines AMD hardware memory encryption, IBM Storage Scale, and Thales CipherTrust for full data lifecycle encryption.
- Multi-factor authentication and OS hardening ensure compliance with GDPR, ISO 27001, and FIPS standards.
- Biomedical benchmarks show security overhead is minimal, preserving high performance for sensitive data workloads.
Why It Matters
Enables life sciences and other fields to process sensitive data at scale without compromising speed or compliance.