Glossary · Performance
Latency: definition, explanation, and why it matters
The round-trip time, measured in milliseconds, for a packet of data to travel from your device to a server and back.
Definition
The round-trip time, measured in milliseconds, for a packet of data to travel from your device to a server and back.
Plain-English explanation
Latency is the delay between asking the network for something and beginning to receive a response. It is measured in milliseconds (ms) and is independent of how much data you can move per second.
When you click a link, the browser sends a request. Latency is the time before the first byte of the reply arrives. Even on a multi-gigabit connection, every interaction begins with this delay — which is why low-latency connections feel snappier than high-bandwidth ones.
Real-world examples
- A 20 ms ping to a regional CDN node feels instant; a 250 ms ping to a distant continent feels sluggish.
- Competitive online games target sub-50 ms latency. Above 100 ms, hit detection and movement start to feel laggy.
- Loading a webpage requires dozens of sequential round-trips; cutting latency in half can noticeably speed up page loads even if bandwidth is unchanged.
Why it matters
Latency dominates user experience for interactive workloads — gaming, video calls, terminal sessions, and any page load that involves multiple requests.
Most 'slow internet' complaints on adequate-bandwidth plans turn out to be latency or jitter problems, not throughput problems.
Common misconceptions
- Myth: More bandwidth fixes high latency.
Reality: Bandwidth and latency are independent. A 10 Gbps link can still have 200 ms ping if the route is physically long or congested. - Myth: Latency is the same as speed.
Reality: 'Speed' usually refers to throughput (Mbps). Latency is delay (ms). A connection can be high-bandwidth and high-latency at the same time.
Frequently asked questions
- What is a good latency value?
- Under 20 ms is excellent, 20–50 ms is great for most uses, 50–100 ms is acceptable for casual browsing, and over 150 ms degrades interactive applications noticeably.
- How is latency different from ping?
- 'Ping' is a tool that measures latency. The latency value reported by ping is the round-trip time. The terms are often used interchangeably in casual conversation.
- Can latency change throughout the day?
- Yes. Network congestion, routing changes, and Wi-Fi interference all add variable latency. Measuring at peak hours gives a more honest picture than off-peak measurements.