Latency

Latency is a measure of how long a transaction takes to complete.

This measure is speed on an atomic transaction level: how fast can a server complete a transaction? Latency is measured in time per transaction. Typical transaction latencies are measured in milliseconds (ms) per transaction.

An analogy might be, how long does it take someone to drive from point A to point B? Typically the trip can be made in 3 hours without traffic; the transaction latency is 3 hours for this trip. However, the average latency for this trip might be 3.5 hours per trip because there is typically traffic congestion that slows the driver down. This analogy illustrates how latency is a representation of an atomic event; the drive from point A to point B.

Latency can be critical for end-user driven workloads.