| Measurement | What it means | What depends on it |
|---|---|---|
| Download speed | How fast you receive data, in megabits per second | Streaming, downloads, page loading, updates |
| Upload speed | How fast you send data | Video calls, sending files, cloud backup, live streaming |
| Ping (latency) | The round-trip time of a packet, in milliseconds | Online gaming, video calls, general responsiveness |
Mind the units: internet plans are sold in megabits per second (Mbps), while download managers often show megabytes per second (MB/s). One byte is 8 bits, so a 400 Mbps connection downloads at about 50 MB/s at best.
Volt relies on M-Lab (Measurement Lab), an open measurement platform founded by researchers and supported by partners including Google. When the test starts, Volt asks M-Lab for the server nearest to you, then uses the ndt7 protocol:
M-Lab servers are connected at 10 Gbps and spread around the world, so the server is not what limits the result, even on very fast fibre. If M-Lab cannot be reached, for example behind a corporate firewall that blocks WebSockets, Volt shows a failure message rather than an approximate figure.
| Use | Recommended download | Recommended upload | Recommended ping |
|---|---|---|---|
| Browsing, email | 5 Mbps | 1 Mbps | under 100 ms |
| HD video streaming | 5 to 10 Mbps per screen | - | - |
| 4K streaming | 25 Mbps per screen | - | - |
| HD video calls | 5 Mbps | 3 Mbps | under 80 ms |
| Competitive online gaming | 25 Mbps | 5 Mbps | under 40 ms |
| Remote work with large files | 100 Mbps | 50 Mbps | - |
For a household, add up simultaneous uses: two 4K screens, a video call and background downloads already need more than 60 Mbps. For gaming and video calls, ping matters more than speed: 1 Gbps fibre with 80 ms of latency will feel worse for gaming than a 50 Mbps connection with 15 ms.
The advertised speed is a theoretical maximum, measured as close as possible to your provider's equipment. Several things explain a gap:
If your speed stays clearly below your plan, over Ethernet and at different times of day, note the results with their dates and contact your provider: regular measurements are a strong argument.
On mobile, speed varies widely with coverage, the frequency band in use and the number of people connected to the same mast. A test indoors can give two to ten times less than outdoors. Remember that the test also uses data: on a fast 5G connection, one test can use several hundred megabytes.
Wi-Fi, other connected devices, the time of day and your computer's limits all affect the result. For a reliable measurement, use Ethernet, close other apps and repeat the test at different times.
Yes. The measurement uses M-Lab servers, connected at 10 Gbps and spread worldwide, and the test picks the one nearest to you. A 2, 5 or 8 Gbps line is measured correctly, as long as your device and cable can keep up.
Volt shows a failure message. This is deliberate: a test hosted on a single small server would only measure that server, and it is better to say so than to display a wrong figure. Try again from another network if a firewall blocks WebSockets.
The test only exchanges random bytes, no files or content of yours. M-Lab records the measurement itself (speed, latency, IP address) and publishes it as open research data. Volt itself stores no results.
It depends on your speed: about 20 seconds of transfer at full speed. On a 100 Mbps connection, expect roughly 150 to 250 MB. Keep that in mind on a limited mobile plan.
Test over Ethernet and close other applications for a reliable measurement.