Which two application requirements are mandatory for traffic to receive proper treatment when placed in the priority queue? (Choose two)

WRED (Weighted Random Early Detection) is a congestion-avoidance mechanism applied to best-effort and class-based queues, not to the priority queue, which uses strict scheduling with a policed rate.

Priority queue traffic must exhibit small, transaction-like behavior (similar to HTTP requests) to prevent queue monopolization - large bulk transfers placed in the priority queue would starve all other queues because the strict priority scheduler services it before all others.

Traffic in the priority queue must be intolerant to packet loss (not tolerant) - real-time applications like VoIP degrade severely when packets are dropped, which is precisely why they receive strict priority treatment.

The fundamental purpose of LLQ is to guarantee bounded latency and jitter for real-time applications like voice and interactive video; only traffic that is intolerant to jitter qualifies, as the queue provides the consistent low-delay forwarding these applications require.

TCP-based applications are poor candidates for the priority queue because TCP's retransmission and windowing behavior can cause bursts that exceed the policed priority queue rate; real-time UDP applications like RTP voice are the intended use case.