Time is a signal
A delay is evidence, not a complete diagnosis.
Round-trip time (RTT) measures the elapsed time between sending a probe and receiving its response. It includes the path in both directions and the time spent processing the probe.
A single sample can be unusual. Repeated measurements help reveal a typical value, variation between samples, and whether some probes receive no response.
Variable delay
Jitter appears when packets wait different amounts of time.
Queues grow and drain as traffic competes for a link. Congestion can make one packet wait behind a burst while the next packet passes sooner. Retransmissions, route changes, and different processing paths can add further variation.
Real-time voice, video, and games are sensitive to this variation because playback and interaction expect a steady arrival rhythm. A jitter buffer can collect packets briefly and play them at a more regular pace, but the buffer adds delay and cannot hide an indefinitely late or missing packet.
Build the model
Four measurements
RTT
The elapsed time for a probe to go out and its response to return.
Median
A typical middle value that is less affected by one extreme sample.
Jitter
Variation in delay between successive measurements.
Packet loss
Probes or packets that receive no response or fail to arrive.
Mini investigation
Read a sample set
Select the observed pattern and identify the strongest first conclusion.
Samples cluster around 24 ms with little variation and no missing responses.
The path appears responsive and consistent for this sample window.
Knowledge check
Test your latency model
Answer eight questions. The answer order changes each time.