A data science team works primarily at a central data center but needs to run a short-term, burst- compute training job in the public cloud to take advantage of specialized GPUs that are not available on- premises. They need to efficiently and securely move a 20 TB dataset from their on- premises ONTAP cluster to a Cloud Volumes ONTAP instance for the duration of the job.
The data flow requirement is as follows:
Source: On-premises ONTAP cluster
Destination: Cloud Volumes ONTAP in AWS
Requirement: Efficient, secure, block-level data transfer.
Which NetApp technology is the most appropriate tool for this task?

• A. NetApp FabricPool
• B. NetApp FlexCache
• C. NetApp SnapMirror
• D. NetApp XCP

An organization has a core data center with a large AI training cluster and several remote edge locations for data ingest and local inference. The edge locations frequently need access to the latest models trained in the core data center, but WAN bandwidth is limited and can be unreliable.
Users at the edge are reporting slow model loading times.
An architect reviews the data access logs from an edge site:
Timestamp: 2025-07-11T15:30:00Z
Event: Model_Load_Request
Model_Path: nfs://core-filer.example.com/vol/models/latest_model.pkl
Source_IP: 192.168.100.15 (Edge Server)
Destination_IP: 10.1.1.50 (Core Filer)
Status: SUCCESS
Duration: 3600s (60 minutes)
What is the most likely cause of the slow model loading times at the edge?

• A. The model file is being transferred over a slow, high-latency WAN link for every load request.
• B. The edge server does not have enough RAM to cache the model effectively.
• C. The NFS version used between the core and edge is outdated.
• D. The core ONTAP filer is using slow, capacity-based disks.

What is the primary architectural advantage of using a NetApp AIPod with NVIDIA DGX servers for the AI training cluster, as described in the scenario?

• A. It is a reference architecture that is pre-validated by NetApp and NVIDIA to eliminate design complexity and ensure predictable performance for AI workloads.
• B. It exclusively uses object storage, which simplifies access for data scientists using S3-native tools.
• C. It is designed for small-scale, departmental AI projects and cannot be scaled.
• D. It prioritizes CPU performance over GPU performance for traditional machine learning algorithms.

What is the primary reason for implementing a hybrid cloud architecture in this medical imaging scenario?

• A. To exclusively use cloud-native AI services, eliminating the need for any on-premises hardware.
• B. To reduce the overall cost of storage by keeping all data in the public cloud.
• C. To leverage the powerful, elastic GPU compute resources available in the public cloud for model training while keeping sensitive source data on-premises to meet security and HIPAA compliance requirements.
• D. To simplify network management by using a single, global virtual network spanning the on- premises and cloud environments.

A data science team reports that their Jupyter notebook pod, which was previously working, is now failing to start. The pod's status is 'CrashLoopBackOff'. An MLOps engineer investigates and finds that the pod's PersistentVolumeClaim (PVC) is bound, but the pod logs show a "Permission denied" error when trying to write to its '/data' mount point.
The engineer checks the Trident backend configuration associated with the pod's StorageClass:
apiVersion: trident.netapp.io/v1
kind: TridentBackendConfig
metadata:
name: ontap-nas-eco
spec:
version: 1
storageDriverName: ontap-nas
managementLIF: 10.10.20.5
dataLIF: 10.10.20.10
svm: svm-prod-ds
exportPolicy: read-only-policy
What is the most likely cause of the "Permission denied" error?

• A. The 'dataLIF' is configured incorrectly and is unreachable from the Kubernetes nodes.
• B. The Kubernetes pod has an invalid 'securityContext' that prevents it from writing to any volume.
• C. The Trident backend is configured to use an export policy ('read-only-policy') that does not grant write permissions to the Kubernetes nodes.
• D. The 'storageDriverName' should be 'ontap-san' for all AI workloads.