Question 370
At fictional Beacon Medical, an HPE edge-to-cloud solution architect supporting healthcare infrastructure is assessing industry architectures and deployment-model tradeoffs during a customer operating model review (HPE-HPE0-V27-TRENDS-093). The reported customer challenge is: A backup design sets a two-hour recovery-time objective but measures only how quickly snapshots are taken. Which repeatable review, test, or governance safeguard would best prevent this problem from recurring?
Explanation
Correct Answer
The best ongoing safeguard is: Conduct timed restoration drills including network and DNS dependencies. It specifically reduces recurrence of: A backup design sets a two-hour recovery-time objective but measures only how quickly snapshots are taken. The relevant correction is: Include application boot, dependencies, identity and user acceptance in end-to-end recovery measurement. Proof of the corrected state is: Business users can complete a validated transaction within the required two-hour interval.
Incorrect Answers
“Include peak login and representative graphics workloads in acceptance tests” supports the separate task “Model concurrency, profile storage, network throughput and GPU needs using representative user personas”, not the reported problem: a backup design sets a two-hour recovery-time objective but measures only how quickly snapshots are taken. This case instead calls for “Include application boot, dependencies, identity and user acceptance in end-to-end recovery measurement”.
“Benchmark end-to-end model pipelines rather than relying on server processor counts” concerns the separate step “Collect model size, accelerator profiles, data loading rate and training concurrency for design”. The reported issue is that a backup design sets a two-hour recovery-time objective but measures only how quickly snapshots are taken, so the direct response is to include application boot, dependencies, identity and user acceptance in end-to-end recovery measurement.
The proposed “Capture workload traffic matrices before selecting switch uplink ratios” tests a different control: derive network fabric topology and oversubscription from actual traffic paths and growth. The reported condition is that a backup design sets a two-hour recovery-time objective but measures only how quickly snapshots are taken; the more direct action is to include application boot, dependencies, identity and user acceptance in end-to-end recovery measurement.