An enterprise needs sub-millisecond forwarding for voice while ensuring voice cannot use more than 1 Mbps of a WAN link. Which configuration provides strict low-latency treatment and limits voice bandwidth?

Configuring LLQ (Low Latency Queuing) with a specified priority bandwidth limit (e.g., 1 Mbps) for the voice class in a service-policy is the correct approach. LLQ provides strict priority queuing for critical traffic like voice, ensuring it is forwarded before other traffic, which is essential for sub-millisecond latency. The priority command with a bandwidth value also caps the amount of bandwidth this class can consume, preventing it from starving other traffic while still guaranteeing its performance. Using CBWFQ without priority queueing would not provide the strict low-latency treatment required. LLQ without a specified bandwidth limit would give voice unlimited priority, potentially starving other traffic, which violates the requirement to limit voice bandwidth. A parent shaper alone, relying on FIFO, would not provide the necessary priority or bandwidth guarantees for voice.