Question 20
A team has corrected this HPE compute solution issue: A university wants live video detection but approves an overnight batch-inference architecture. Which continuing practice would best prevent the same problem?
Explanation
Correct Answer
Record timing and throughput requirements before selecting inference modes. This recurring assurance measure is tied to the diagnosed risk: Batch jobs cannot satisfy a real-time camera alert latency objective. For the active incident, the corrective action remains: Use a supported streaming or low-latency inference pipeline matched to camera deadlines.
Incorrect Answers
Batch jobs cannot satisfy a real-time camera alert latency objective. The proposed safeguard is intended for another failure mechanism: Model modalities and task capabilities vary and must be explicitly verified. Record timing and throughput requirements before selecting inference modes.
Supervised training needs reliable labels and a representative reference dataset. The option addresses that risk, not the one responsible for the present issue. Batch jobs cannot satisfy a real-time camera alert latency objective. Record timing and throughput requirements before selecting inference modes.
The prevention measure is relevant to this condition: A convenient inference endpoint may violate data-residency and confidentiality requirements. Here the recurring weakness is: Batch jobs cannot satisfy a real-time camera alert latency objective. Record timing and throughput requirements before selecting inference modes.