An airport operations data center must redesign CRAC and airflow because rear-door heat from GPUs overwhelms older rows. Which action addresses the root issue?

An airport operations data center overwhelmed by GPU rear-door heat in older rows must redesign cooling and airflow—CRAC capacity, containment, and row airflow—for the new heat loads. The root issue is thermal design mismatch, not simply increasing cold-aisle perforated tile airflow for GPU rear-door heat loads, which fails without CRAC and containment upgrades. Domain 2.4 requires cooling/airflow redesign as part of AI densification when legacy rows cannot reject GPU heat.

Increasing cold-aisle perforated tile airflow for GPU rear-door heat loads fails by concept: it does not address CRAC and containment capacity for rear-door heat, only mixing hot air. Fitting rear-door heat exchangers to legacy chilled water loop for GPU heat loads fails safety and operations—insufficient capacity is not remediation. Upgrading hot-aisle return fans (CRAC/containment) for GPU rear-door heat loads fails Domain 2.4; return fans alone do not fix the root cooling mismatch.

Exam caveat: Redesign may lead to containment, CRAC replacement, or liquid-assist options in follow-on decisions. This item's correct action is addressing airflow/CRAC for rear-door GPU heat, not upgrading hot-aisle return fans (CRAC/containment) for GPU rear-door heat loads or postponing facility work indefinitely.

Operational check: do not fit rear-door heat exchangers to legacy chilled water loop for GPU heat loads; map hot spots, model rear-door heat versus CRAC/airflow capacity, implement containment or CRAC upgrades and re-validate under airport training load before further GPU adds in those rows.