Network latency

Network latency is the time it takes data to travel from a source to a destination across a network, commonly measured as round-trip time (RTT), the time for a packet to reach its destination and a response to return. It is typically reported in milliseconds and measured with tools like ping or embedded directly in application-level monitoring.

Latency differs from packet loss and from throughput: a connection can have low latency but still drop packets under congestion, and a high-bandwidth link can still have high latency if the physical distance or number of hops is large, since latency is bounded by physical distance and routing path in a way raw bandwidth is not. Applications sensitive to latency, video calls, gaming, financial trading, are affected differently than those sensitive mainly to throughput, like bulk file transfer.

Latency matters because it directly shapes user-perceived performance and is a core input to quality of experience and application performance monitoring measurements. A common pitfall is monitoring only average latency, which hides the tail: a small share of very slow requests can dominate user complaints even while the average looks acceptable, so percentile measurements matter more than the mean.