FHE-based platform for secure analytics —
never expose raw information
Public agencies, healthcare institutions, and financial institutions hold some of the most valuable data in existence — yet the privacy risks and regulatory burden involved in unlocking that value have long gone unsolved.
DESILO DCR performs both statistical and machine-learning analyses directly on encrypted data, ensuring the raw records are never exposed in plaintext — not even once during the entire analysis pipeline. It delivers a true homomorphic-encryption-based data clean room to the public, healthcare, and financial sectors.
![[ System diagram — designer to provide ] · DESILO DCR analysis flow (upload → analysis on encrypted data → multi-party distributed decryption)](/images/solutions/ddcr-diagram-mobile-en.png)
From Data Assetization to Real-World Analysis
Statistics and machine learning, all on encrypted data
Encrypted statistical and ML analysis
Performs standard statistical and machine-learning analyses — frequency analysis, regression, classification — directly on uploaded data, without decryption.
User-defined function (UDF) support
Beyond predefined analysis functions, analysts can run their own UDFs in the homomorphic-encryption environment as-is — preserving full analytical flexibility.
Multi-party key management and distributed decryption
Decryption authority is never granted to any single party. Results are revealed only when a pre-agreed quorum of authorized parties jointly consents.
Bidirectional deployment: SaaS and on-premises
Use it as a SaaS offering on public cloud, or install on-premises in environments with strict network isolation or security requirements.
FHE-based Federated Learning
Supports model training directly within each participant's data environment — without ever collecting or centralizing raw data.
FHE applied throughout the training process reduces the privacy risks of intermediate information sharing, delivering a secure multi-institution collaborative training environment that preserves data confidentiality.
A proven, procurable data-analysis infrastructure
Designated as a Korean Public Procurement Service Innovative Product
Designated as an Innovative Product by the Ministry of Science and ICT and the Public Procurement Service in 2025, enabling pilot procurement and direct-contract procurement by public agencies.
Fully qualified for the public-sector market
Certified by the Korea Internet & Security Agency (KISA) with CSAP SaaS Standard certification (CSAP-2026-015) and holds GS Certification Grade 1 — completing the qualifications to supply SaaS services to public agencies.
Cleared the Personal Information Protection Commission's prior-adequacy review
Cleared the Personal Information Protection Commission's prior-adequacy review in January 2024, minimizing the compliance burden on adopting institutions.
Deployed at the National Cancer Center
Deployed in the National Cancer Center's Personal Information Innovation Zone and currently in production as a real-world patient-data analysis environment.
Industry-leading performance powered by DESILO FHE Library
Backed by a best-in-class proprietary homomorphic encryption library, delivering processing speeds that decisively outpace other FHE-based analytics solutions.
A data-analysis environment already running in the public, healthcare, and financial sectors
Conduct clinical research, epidemiology, and AI-model training safely on EMR and genomic data — without the data ever leaving the institution. In production at the National Cancer Center's Personal Information Innovation Zone.
Applied to fraud detection and credit-scoring model training and validation on sensitive financial data such as transaction histories and credit information.
Run policy-impact analysis, gap detection, and statistical reporting on sensitive administrative data such as residency, welfare, and statistics records — without security overhead.
Use as an analytics infrastructure for hard-to-share data such as national statistics, census, and R&D datasets.