Latency, defined
Latency is the time it takes for a small piece of data — a packet — to travel from your device to a destination and back again. That there-and-back measurement is called the round-trip time, or RTT, and it is what the familiar ping command reports. It is measured in milliseconds (ms), thousandths of a second.
When people say a connection feels “snappy” or “laggy,” they are almost always describing latency, not bandwidth. Every click, keystroke in a remote session, or shot fired in an online game has to make at least one round trip before you see the result. Low latency means the result comes back almost immediately; high latency means you wait, however briefly, every single time.
What causes latency
A packet's round-trip time is the sum of several delays stacked on top of each other. No single one dominates in every case, which is why latency varies so much between connections.
- Physical distance and the speed of light. Signals in fiber travel at roughly two-thirds the speed of light in a vacuum. That sounds instant, but it is a hard floor: a round trip across a continent or an ocean simply cannot happen faster than light allows. A server thousands of kilometres away will always add tens of milliseconds no matter how good your connection is.
- The number of hops. Your packets pass through many routers on the way to their destination. Each hop adds a small processing and forwarding delay, and a longer or less direct path means more hops.
- Queuing and congestion. When a link is busy, packets wait in buffers before they can be sent. Under load this queuing delay can dwarf everything else — the effect known as bufferbloat, where a saturated upload makes your ping spike dramatically.
- Last-mile technology. The access technology between you and your ISP matters. Fiber and modern cable add very little; older DSL, cellular radio, and especially satellite add their own overhead before your packet even reaches the wider internet.
- Serialization. It takes a moment to clock the bits of a packet onto the wire. On fast links this is negligible, but on slow links it becomes a measurable part of the delay.
Latency vs. bandwidth: a fat pipe does not make a far server closer
This is the single most useful thing to understand about networking performance. Bandwidth is capacity — how much data can move per second, quoted in megabits per second. Latency is delay — how long one packet takes to make the round trip. They are almost independent of each other.
The classic analogy is a highway. Bandwidth is the number of lanes; latency is how long the drive takes. Adding lanes lets more cars travel at once, but it does not make any individual car arrive sooner. Upgrading from a 100 Mbps plan to a 1 Gbps plan widens the road — great for large downloads and many simultaneous streams — but it does nothing for the drive time to a distant server. A connection with a huge download speed and high latency still feels laggy for anything interactive.
This is why gamers, video callers, and remote-desktop users care far more about ping than about their advertised speed tier. Once you have enough bandwidth to carry your traffic, extra bandwidth stops helping and latency becomes the thing you feel.
Jitter: the variation that wrecks calls and games
Jitter is the variation in latency from one packet to the next. If your ping is a steady 30 ms, jitter is low. If it bounces between 20 ms and 120 ms, jitter is high — even though the average might look fine on a summary screen.
Real-time applications hate jitter more than they hate a steady delay. A video-call app can adapt to a consistent lag by keeping a small buffer, but uneven arrival times break that assumption. Packets that show up late get dropped or force the buffer to grow, which you experience as choppy audio, robotic voices, frozen frames, and stutter in games. A link with slightly higher but rock-steady latency will usually feel better than one with a lower average that swings wildly.
Typical ranges (approximate)
Real numbers depend heavily on how far away the thing you are measuring is, so treat these as rough, qualitative ballparks for reaching a nearby, well-connected server — not precise figures.
- Wired fiber, local: single-digit to low-tens of milliseconds. This is about as good as consumer connections get.
- Cable and DSL: a bit higher than fiber, often low-tens of milliseconds, with DSL and older gear tending toward the upper end.
- LTE and 5G: higher again, because the radio link adds overhead. Good 5G can approach wired figures; typical mobile is noticeably above fiber.
- Geostationary satellite: very high — around 600 ms round trip is common, because the signal has to travel roughly 36,000 km up to the satellite and back down for each leg. Physics sets this floor.
- Low-Earth-orbit satellite (such as Starlink): much lower than geostationary because the satellites orbit far closer to the ground, bringing latency down into a range that is usable for interactive traffic, though still above good wired connections.
How to measure it
You do not have to guess. The built-in IPFerret speed test measures latency and jitter alongside download and upload throughput, so you can see all four numbers together rather than just the headline speed. Run it a few times and watch whether the latency and jitter stay stable or jump around.
To find out where the delay is accumulating, use a traceroute. It lists every router hop between you and a destination and shows the round-trip time to each one, so you can spot the exact point where latency jumps — whether that is your own last mile, a congested peering link, or simply the long haul to a distant server.
How to reduce latency
Some latency is fixed by physics — you cannot beat the speed of light to a faraway server. But a surprising amount is within your control.
- Use a wire instead of Wi-Fi. Wireless adds variable delay and jitter, especially on a congested channel or at the edge of range. A direct Ethernet run is the single biggest easy win for responsiveness; see Wi-Fi vs. Ethernet for the trade-offs.
- Connect to closer servers and CDNs. Distance is latency. Choosing a game server, voice region, or download mirror near you cuts the round trip directly. A content delivery network does this automatically by serving you from a nearby edge location instead of a distant origin.
- Avoid congestion. A saturated link buffers packets and spikes your ping. Do not run large uploads or downloads while gaming or on a call, and consider hardware that manages bufferbloat.
- Enable QoS. Quality-of-Service settings on many routers prioritise latency-sensitive traffic — voice, video, gaming — over bulk transfers, keeping your ping steady even when the connection is otherwise busy.
- Prefer a modern access technology. If you have a choice, fiber or good cable will nearly always beat DSL, cellular, or satellite for latency. A dual-stack IPv6 connection can also help by avoiding extra translation layers on some networks.
The bottom line
Bandwidth gets all the marketing attention because it is a single big number that is easy to advertise. But for everything interactive — calls, games, remote work, even the feel of a website loading — latency is what you actually experience, and jitter is what turns a good connection into a frustrating one. Measure all of them, not just the download figure, before you conclude your connection is fast.
Frequently asked questions
What is a good ping?
It depends on what you are doing, but as a rough guide: under about 20 ms feels instant and is excellent for competitive gaming and video calls; roughly 20 to 60 ms is good and fine for almost everything; 60 to 100 ms is usable but noticeable in fast-paced games; and above 100 ms starts to feel laggy for real-time applications. For web browsing and streaming, even a few hundred milliseconds is tolerable because those workloads are not interactive frame-by-frame. The numbers are approximate and vary with the destination you are measuring against.
Is latency the same as internet speed?
No. Speed usually means bandwidth — how much data can move per second, measured in megabits per second. Latency is how long a single packet takes to make the round trip, measured in milliseconds. They are independent: you can have a very fast connection with high latency (a satellite link) or a modest connection with very low latency (local fiber). A bigger pipe does not make a distant server respond any sooner, which is why a high download number does not fix a laggy feel.
Why does jitter matter more than raw latency for calls and gaming?
Jitter is the variation in latency between packets. Real-time applications can compensate for a steady delay, but they struggle when packets arrive unevenly. High jitter forces a video-call app to grow its buffer or drop late packets, which shows up as choppy audio, frozen frames, and stutter. A connection with a slightly higher but stable latency often feels better than one with a lower average latency that swings wildly.
Related reading
- Speed test — measure your latency, jitter, download, and upload in one run.
- Traceroute — see exactly which hop is adding the delay.
- Wi-Fi vs. Ethernet — why a wire is the easiest latency win.
- What is a CDN? — how serving from nearby edges cuts round-trip time.
- IPv6 status — check whether your connection has working IPv6.
- Networking glossary — plain-English definitions of ping, RTT, and more.
