FRIDGE is a DARE UK Early Adopter project that extends Trusted Research Environments (TREs) to securely leverage the computational resources of the national AI Research Resource (AIRR). It provides a framework for submitting and managing sensitive data workflows from within a TRE to an external FRIDGE cluster, while enforcing data governance and security policies throughout. The system is built on Kubernetes for portability across cloud providers and on-premise installations.
Full documentation is available here
| Directory | Description |
|---|---|
fridge-job-api/ |
FastAPI service for submitting and managing workflows, and passing data between the FRIDGE and TRE |
infra/ |
Pulumi infrastructure-as-code for supported deployment targets |
docs/ |
Source for the project documentation site |