What Is Latency?

Latency and speed answer different questions

Speed describes capacity: how much data your connection can move in a given second, usually reported in megabits per second. Latency describes responsiveness: how long a single message takes to get to its destination and back. A connection can excel at one and struggle at the other, because they depend on different things.

Latency

The time delay between sending a request and receiving a response, usually reported as round-trip time in milliseconds.

Downloading a large file mostly depends on speed — more capacity gets more data through per second. Loading a web page, joining a video call, or reacting to something in an online game depends heavily on latency, because those activities involve many small exchanges where each one has to complete before the next step can happen. A very fast connection with high latency can still feel sluggish for exactly this kind of use, even while a large file download on the same connection completes quickly.

Round-trip time, and why it is measured both ways

Latency is almost always reported as round-trip time: the time from sending a small probe to receiving a reply, rather than the one-way delay in either direction alone. This is partly a practical choice — measuring a one-way delay accurately requires perfectly synchronised clocks at both ends, which round-trip measurement avoids — and partly because most everyday activity, like loading a page or reacting in a call, depends on a full round trip rather than a single direction.

A single round-trip measurement is a snapshot, and results vary from one probe to the next even on a healthy connection. This is why latency is normally reported alongside a related but distinct measurement — jitter — which describes how much that round-trip time varies from one probe to the next, rather than relying on one number alone.

Idle latency versus the response time you actually experience

Idle latency is measured with the connection otherwise doing nothing else, which represents close to the best case a connection can offer. In daily use, a connection is rarely completely idle — something is usually syncing, updating, or streaming in the background — and the response time you experience in practice can be somewhat higher than the idle figure, particularly while other traffic is competing for the same link.

It is worth being cautious about treating idle latency as a fixed minimum that always applies. It is closer to a floor: the lowest delay you would expect to see under favourable conditions, rather than a guarantee of what every interaction will feel like. The gap between idle latency and the delay experienced while the connection is busy is covered separately, since it depends on how queues on the path are managed rather than on distance alone.

Physical distance and routing both add delay

Signals travel through cables and equipment at a fraction below the speed of light, and every additional kilometre a signal has to travel adds a small, unavoidable amount of delay. A server on the other side of the world will always have higher latency than one in the next city, no matter how good the connection to either one is, simply because of the distance involved.

Distance is only part of the picture, though. Data rarely travels in a straight line — it passes through a series of intermediate networks and routing equipment, and the specific path chosen is determined by agreements between network operators rather than by geography alone. Two destinations at similar physical distances can have noticeably different latency because the routing path to one happens to be more direct than the path to the other.

This also explains why latency to different destinations from the same connection can vary quite a bit. A gaming server, a video call service, and a general website may each be reached by a different path, with a different number of hops and a different total distance travelled.

Wi-Fi, mobile networks and VPNs each add their own delay

The internet connection itself is only part of the path between your device and a destination. A wireless link between your device and your router adds its own small delay, generally larger and more variable than a wired connection, particularly when the signal is weak or the wireless channel is shared with many other devices nearby.

Mobile networks add delay in their own way, since a signal has to travel to a cell tower and be processed by the mobile operator's network before it ever reaches the wider internet. This structure means mobile latency is often higher and more variable than a comparable fixed-line connection, even when download speed looks similar.

A VPN adds delay for a different reason: your traffic is deliberately routed through the VPN provider's server before continuing to its actual destination, which almost always makes the physical and routing path longer than it would be otherwise. This is an expected consequence of how a VPN works, not typically a sign of a fault.

• Wi-Fi: adds variable delay, worse with a weak signal or heavy channel congestion.
• Mobile networks: adds delay from the path between your device, the cell tower, and the operator's core network.
• VPNs: adds delay by design, since traffic is routed through an additional server before reaching its destination.

What latency actually changes in daily use

High latency makes interactive activities feel delayed even when nothing is technically wrong with the connection's capacity. In a video call, high latency can create an awkward lag where people start talking over each other because responses take a fraction of a second longer to arrive than expected. In gaming, latency directly affects how quickly an action is registered by a game server, which is why players often refer to it by the more familiar term 'ping'.

Even ordinary browsing is affected: loading a page usually involves dozens of small back-and-forth exchanges to fetch different pieces of content, and each one is delayed by the connection's latency. A page with high latency to its server can feel slow to start loading even on a connection with plenty of download speed available, because the bottleneck is the round trip, not the bandwidth.

Why low speed and high latency are separate problems

It is easy to assume that a connection which struggles feels slow for one single reason, but low download speed and high latency usually come from different parts of the connection and are fixed in different ways. Low download speed generally reflects the plan itself or the capacity of the link into your home. High latency generally reflects distance, routing, or an added hop like Wi-Fi, mobile access, or a VPN.

A connection can have very high download speed and still have high latency to a specific destination because of where that destination is located and how it is reached. Equally, a connection with modest download speed can have very low latency if the path to the destination is short and direct. Treating the two as the same problem can lead to the wrong fix — upgrading a plan will not reduce the physical distance a signal has to travel.

What Norynt measures, and its limitations

Norynt measures idle latency as a round-trip time to its test infrastructure, alongside download speed, upload speed, jitter, loaded latency, and how steady the speed stays. Idle latency is collected before any transfer starts, giving a baseline reading of how quickly the connection responds when it is not busy.

Norynt measures latency as round-trip time to its own test servers, which is a good general indicator but will not exactly match latency to a specific game server, call service, or website, since each of those may be reached by a different path.

It is also worth being clear about what Norynt does not measure. It does not measure packet loss directly, so a latency or jitter result should not be read as evidence about whether packets are being dropped along the path. Latency describes delay; it does not by itself describe loss.