Why is quality of service critical to voice network traffic?

Voice traffic is real-time, meaning packets must arrive within strict delay budgets (typically under 150 ms one-way) and with minimal jitter to be intelligible. Unlike data traffic, lost or late voice packets cannot be retransmitted because doing so would introduce unacceptable delay that disrupts the conversation. QoS mechanisms like prioritization queuing (LLQ/CBWFQ) and traffic marking (DSCP EF) ensure voice receives preferential treatment to meet these real-time requirements.

Voice uses UDP and does not support retransmission; resending packets would introduce latency far exceeding acceptable thresholds and would actually degrade call quality.

Voice traffic is not bursty - it produces a consistent, predictable stream of small packets at fixed intervals (e.g., every 20 ms), unlike data applications that generate traffic in bursts.

Voice traffic can be compressed using low-bitrate codecs such as G.729 and G.726, so the inability to compress is not a valid reason why QoS is necessary.