Throughput
How much work or data a system can handle over a given stretch of time.
Glossary
Latency
The delay between asking for something and the moment the answer begins to arrive.
In one line
The delay between asking for something and the moment the answer begins to arrive.
Also calleddelayresponse time
Latency is about waiting. It measures how long it takes for a response to start, regardless of how much data eventually follows. Low latency feels immediate; high latency feels sluggish, even when everything is otherwise working perfectly.
Much of the quiet craft in a driver is about keeping latency low, so that interactions feel instant even when a great deal is happening underneath. A small delay repeated many times can add up to a noticeable drag, so trimming each one matters.
Latency and the amount of data moved are separate qualities. A system can begin responding almost instantly yet deliver slowly, or take a moment to start yet then move a great deal at once. Understanding which one you are feeling helps explain why something seems fast or slow.
Some latency is unavoidable, set by physical distance and the time signals take to travel. The goal is rarely to remove it entirely but to keep it small and steady, because a delay that is short and predictable is far easier to live with than one that jumps around.
In practice
In practice, latency is like the pause between flipping a switch and the light coming on. If the light responds the instant you touch the switch, it feels effortless; if there is a noticeable gap, it feels broken even when the bulb is fine. Software aims for that instant response.
