A hospital branch prioritizes voice with strict 256-kbps priority and no queuing delay, while file backup in the same VPN needs a guaranteed minimum share during congestion but can wait. Both classes share a 10-Mbps tunnel where voice must never starve data. Which WAN Edge scheduling and queue class combination meets this voice latency guarantee?

The correct combination works because low-latency queuing gives voice strict priority up to the 256-kbps policer for no queuing delay and bounded jitter, the policer caps priority traffic so the shared 10-Mbps tunnel cannot be starved by voice bursts, and a bandwidth percent queue gives backup a guaranteed minimum rate during congestion while allowing it to wait. The variant that omits the policer fails on starvation protection because unconstrained priority queuing can monopolize the scheduler if voice exceeds its profile, violating the requirement that voice must never starve data despite correct latency handling. The variant that uses bandwidth percent for both classes fails on scheduling because weighted fair sharing provides only a proportional minimum and still queues voice, so even a high share cannot deliver strict priority or delay bounds. The variant that keeps policed priority for voice but leaves backup in default best-effort fails on assurance because best-effort has no reservation and uses only leftover capacity, so backup loses its required minimum during sustained congestion. As an exam caveat, link strict and no-delay language to policed low-latency queuing and link guaranteed minimum but can wait to bandwidth percent, since large bandwidth alone never equals priority. As a practical check, verify the edge QoS map for queue types and policer rate, then test under load to confirm low voice jitter with exceed drops and measurable backup throughput.