The desired number of model units that was requested to be available for this provisioned throughput.