This is the average latency computed across all sampled test frames at this frame size, giving the typical delay figure most commonly quoted when comparing measured performance against an SLA's latency target. An admin uses this as the primary number for day-to-day SLA verification, reserving the min/max fields for diagnosing variability once the average itself looks acceptable or unacceptable. For example, an average of 120 microseconds well under a 500-microsecond SLA target would let the admin sign off on the latency portion of a service acceptance test.