Deploying a FRIDGE

Deploying a FRIDGE is a multi-stage process.

1. Deploy the networking infrastructure and Kubernetes clusters

2. Deploy FRIDGE services into the access cluster

3. Deploy FRIDGE services into the isolated cluster

4. Perform a final networking lockdown

The Hosting Provider Administrators are responsible for the first and final steps.

The TRE Administrators are responsible for the second and third steps.

For Hosting Provider Administrators, follow the guide in Deploy Infrastructure

For TRE Administrators, follow the guide in Deploy Services

Prerequisites¶

You will need the following tools installed to deploy FRIDGE:

• Python 3.11 or later

• Pulumi

• kubectl

• NetBird

Additionally, if deploying to Azure, you will need the following:

• Azure CLI

NetBird¶

Access to a deployed FRIDGE, either for TRE Administrators or TRE Users, is through a Virtual Private Network managed using NetBird

You will require access to a NetBird Management server. For testing and development, the free NetBird cloud management server is sufficient. However, for production, we recommend using self-hosted NetBird.

For instructions on configuring NetBird, see Configuring NetBird

Pulumi Backend¶

Pulumi stores state in a backend. The Pulumi documentation details how to set up an appropriate backend. For local development and testing, you can use the local backend:

pulumi login --local

For production, another backend, such as Azure Blob Storage, will likely be more appropriate.