Latency

Origin domain

Computer Science & Software Engineering

Also from

Neuroscience, Economics & Finance

Aliases

Delay, Transit Time, Response Delay, Dead Time, Lead Time

Core Idea

Latency is the irreducible time interval between a stimulus entering a system and the corresponding response becoming observable at the output. It is the transit cost of a single signal through a channel, processor, or pathway — distinct from throughput (how much flows per unit time) and from rate mismatch. Where there is latency, what arrives now reflects what happened earlier, so the present output is always a delayed image of a past input.

How would you explain it like I'm…

The Wait

Latency is how long it takes for something to get from one place to another. When you yell into a big canyon, you have to wait a moment before the echo comes back — that wait is latency. It's not about how loud you yell or how much you say, just how long the trip takes.

Delay Time

Latency is the delay between when something starts and when you see the result. If you press a button on a video game and your character jumps half a second later, that half second is latency. It's different from how much stuff a system can handle at once — a wide highway can carry many cars but each car might still take a long time to get across town. Whatever you see now actually happened a little while ago.

Latency

Latency is the time gap between a signal entering a system and the matching response showing up at the output. It's the transit cost for a single signal, and it's different from throughput, which measures how much can flow per unit time. A network can move huge amounts of data per second and still take a long time to deliver any one packet, because delay depends on distance, propagation speed, queuing, and how many stages the signal must cross. The same idea shows up as reaction time in nerves, dead time in control systems, and lead time in supply chains. Whenever there's latency, what you see now is a delayed image of an earlier cause.

 

Latency is the irreducible time interval between a stimulus entering a system and its corresponding response appearing at the output — the transit cost of a single signal through a channel, processor, or pathway. It is structurally distinct from throughput (volume per unit time) and from rate mismatch (a difference in the speeds of coupled processes). Crucially, latency is a property of traversal, not processing capacity: a channel can have enormous bandwidth and still impose long per-signal delay because transit time is set by path length, propagation speed, queueing, and the number of intermediate stages — not by channel width. The concept gained precision in packet networking, where round-trip time is measured independently of throughput, but it generalizes to neuroscience (reaction time), control engineering (dead time, the delay before a controller's action shows up in the plant), supply-chain operations (lead time), and macroeconomics (monetary policy transmission lag). Latency names the gap between cause and observed effect, and that gap is the structural root of an entire family of timing pathologies — overshoot, oscillation, and acting on stale information.

Broad Use

• Computer science / networking: the round-trip delay before a packet or query returns, independent of bandwidth.
• Neuroscience: the conduction and synaptic delay between a sensory event and a behavioral response (reaction time).
• Supply chain: lead time — the gap between placing an order and receiving goods.
• Control theory: dead time between actuating a control and seeing its effect, a notorious source of instability.
• Economics: transmission lag between a central-bank rate change and its effect on inflation.
• Telecommunications / live media: the delay that makes a satellite conversation talk over itself.

Clarity

Naming latency separates how fast a response comes back from how much can be processed. It lets practitioners say "the channel is fast but high-latency" — high bandwidth, slow round trip — a distinction invisible if one only tracks volume. It also makes explicit that any observed state is a past state.

Manages Complexity

Latency bounds reasoning about responsiveness to a single number with units of time, independent of the internal machinery. It lets a designer treat a complex subsystem as a black box characterized by its delay, and it concentrates instability analysis on where in a loop the delays accumulate.

Abstract Reasoning

Recognizing latency enables a key inference: a feedback loop with significant delay between action and observed effect will tend to overshoot, oscillate, or chase stale information. It predicts that decisions made on delayed signals are decisions about the past, and that shortening the loop (not enlarging it) is the lever for stability.

Knowledge Transfer

The control engineer's hard-won lesson — that dead time, not gain, destabilizes a thermostat — transfers directly to why a manager reacting to last quarter's metrics over-corrects, and to why a high-latency network connection makes interactive control jerky. The supply-chain notion of lead time maps onto the brain's reaction-time budget.

Example

A driver adjusting a shower's temperature feels the consequences of each twist only seconds later; the delay tempts them to over-twist, swinging between scalding and freezing. The identical structure governs a high-frequency trader on a slow link acting on prices that have already moved, and a feedback controller whose dead time forces it to act conservatively to avoid runaway oscillation.

Relationships to Other Primes

Parents (2) — more general patterns this builds on

• Latency is a kind of Constraint — Latency is a specific kind of constraint, binding system response time below an irreducible transit minimum.
• Latency presupposes Time — Latency presupposes time because the irreducible interval between stimulus and response is a temporal duration measured along time's ordering.

Path to root: Latency → Constraint

Not to Be Confused With

• Latency is not institutional lag or culture lag, which describe differential rates of change between coupled components producing maladjustment; latency is the simple transit delay of one signal.
• It is not buffering, which is maintained capacity that absorbs variation; latency is the time cost of traversal, which buffering can even increase.
• It is not intermittency, which concerns the bursty timing pattern of events, not the delay between an input and its response.