Send to a cloud archive on Azure or Google Cloud

Goal: every study also goes to a cloud DICOM service, over DICOMweb, without slowing anything else down.

Azure Health Data Services

1. In Azure, create the DICOM service, and an app registration with a client secret. Give the app the DICOM Data Owner role on the DICOM service.
2. In Routes, Configuration › Destinations › Add destination, type DICOMweb:
   • Service URL: https://<workspace>-<service>.dicom.azurehealthcareapis.com/v2;
   • Authentication: OAuth 2.0 client credentials;
   • User name / client ID and Password / token / client secret: the app’s;
   • Azure Health Data Services: your tenant ID, then Fill in for Azure (sets the token endpoint and scope).
3. Save, then Check: it searches for one study with those credentials.

Google Cloud Healthcare API

1. In Google Cloud, create a dataset and a DICOM store, and a service account with the Healthcare DICOM Editor role on the store. Create a JSON key for it.
2. In Routes, add a DICOMweb destination:
   • Service URL: https://healthcare.googleapis.com/v1/projects/<project>/locations/<region>/datasets/<dataset>/dicomStores/<store>/dicomWeb;
   • Authentication: Google Cloud service account;
   • Service account key (JSON): paste the key, or choose the file. It is checked when you save.
3. Save, then Check.

Then

• Add the destination to the routes whose studies should go to the cloud.
• Instances per request (default 10) puts several in one request: far fewer round trips to a distant service.
• Keep the internet line free for the working day with a schedule: a limit in office hours, full speed at night.
• For research or AI use in the cloud, give the destination a de-identification profile.

The secret or key is stored encrypted, never shown again, and given only to the nodes. See DICOMweb.