A library consortium DC tests failing over non-sensitive inference to cloud during civic DC maintenance. What hybrid capability is primarily under evaluation?
Select an answer to reveal the explanation.
Short Explanation
Maintenance windows should not equal total AI blackout for non-sensitive inference. Failing that traffic over to cloud is continuity via workload mobility. Don't treat it as a permanent migration since you'll need to fail it back after maintenance.
Full Explanation
A library consortium DC that tests failing over non-sensitive inference to cloud during civic DC maintenance is primarily evaluating workload mobility that sustains non-sensitive inference in cloud during on-prem maintenance. Continuity for eligible workloads is the point: keep serving while the municipal hall is down for work, without treating every dataset as cloud-eligible.
Backup replication that archives non-sensitive inference to cloud during on-prem maintenance confuses backup with failover; it copies data but does not move inference workloads to sustain service during the outage. Permanent migration that relocates non-sensitive inference to cloud during on-prem maintenance tests a one-way move, not temporary failover; it does not evaluate sustaining service in cloud with failback after the outage. Unified observability that monitors non-sensitive inference in cloud during on-prem maintenance adds observability but does not test workload mobility or keep inference running.
Exam caveat: Domain 2.5 evaluates mobility for operational resilience; only workloads cleared by sensitivity and policy should fail over—backup replication that archives non-sensitive inference to cloud during on-prem maintenance is a backup task, not an inference failover, and stays outside the mobility drill.
Operational check: during a maintenance drill, shift only labeled non-sensitive inference to cloud, confirm backup replication that archives non-sensitive inference to cloud during on-prem maintenance stays a backup task and does not substitute for inference failover, and measure that eligible users still receive inference. Document the eligibility matrix before the maintenance window: which library inference services may move, which stay pinned on-prem, and how DNS or traffic steering returns them after the civic DC is healthy. Mobility drills that skip failback validation leave the consortium stuck on cloud longer than policy intended.
Prefer to study offline? Printable PDF study book + test engine ($7.50, unofficial). These questions stay free here.