This page provides information about data residency with AML AI. Data residency describes where your data is stored at rest. To help comply with your internal and external data residency requirements, AML AI gives you the ability to control where data is stored.
Each Instance resource is created with a region specified. The artifacts created underneath the parent are constrained to that region location.
AML AI and data residency
- Customer Data is stored at rest and machine learning processing is performed in the specified region.
- This includes Customer Data associated with AML AI resources, such as datasets, engine configs, models, and includes copies of inputs and outputs, generated ML features, model hyperparameters, model weights, and prediction results.
- See the Service Specific Terms for the definition of Customer Data, which may not include resource identifiers, attributes, or other data labels.
Organizational constraints
Organizational constraints (constraints/gcp.resourceLocations) can control the
region in which Google Cloud resources can be created. AML AI
obeys these regional constraints at resource creation time. For more information,
see restricting resource locations
and resource locations supported services.
Location resource
AML AI has a Location resource as a parent object of all other resources in AML AI. All child AML AI resources are associated with only one Location resource, which ensures data residency within the associated Google Cloud region.
Input and output data
An AML AI instance requires all input and output data to be specified in the same region and project. This prevents data from being accidentally moved between Google Cloud regions.
What's next
For more information about Google's data residency commitment, read the data residency terms.