A school district data center must expand storage with new shelves while training jobs continue. Which scalability attribute is required?

A school district data center that must add storage shelves while training jobs continue requires online capacity expansion—growth that does not force stopping active AI training. Scalable AI storage platforms add shelves or nodes with services online so academic calendars and overnight district jobs are not parked for every capacity bump. Continuity during expansion is the scalability attribute Domain 2.3 is testing here.

Scale-out cache expansion that adds shelves while sustaining active AI training fails by concept: it addresses cache acceleration, not the online usable capacity expansion the district needs while training continues. Controller IOPS upgrade that tunes new shelves while sustaining active AI training fails scalability and economics here—IOPS tuning is not online storage capacity expansion. Automated cold tiering that vacates shelves while sustaining active AI training fails capacity expansion here—vacating existing shelves is not the same as adding storage shelves online while training jobs keep running.

Exam caveat: Online expansion refers to storage service continuity; firmware or data-motion caveats may still apply, but the exam contrast is online capacity growth versus cache expansion, IOPS tuning, or cold tiering that sustains training but does not add usable capacity online.

Operational check: validate vendor online shelf/node add procedures in a lab; expand before school-year ingest peaks; confirm trainers keep reading/writing through the expansion and that rebuild/rebalance does not push latency past district GPU efficiency targets.