Working with large devices with many models creates unique scale, performance and validation issues. Today, validation of models is strictly enforced, which prevents a device from being mounted if the validation fails for some models. Also, schema validation for a large number of models on a large number of devices can take a long time. Many times, an operator only needs a subset of models from the device.

A solution is required where the operator will be able to define a policy that specifies the subset of models that should be validated for a particular device. This policy then could be applied in a particular mountpoint's configuration. The policy could also specify type of enforcement - strict or loose. In a strict enforcement, mounting would fail if a validation of any model fails. In a loose enforcement, models that fail validation would be ignored and flagged.