Download time:
bits vs bytes, and why 100 Mbps is not 100 MB/s

The factor of eight

Broadband and mobile plans quote speed in megabits per second, written Mbps or Mb/s, with a lowercase b. File sizes, browsers' progress bars and storage devices use megabytes, MB, with an uppercase B. One byte is eight bits, so a 100 Mbps connection moves at most 12.5 megabytes every second.

When your browser shows 11 MB/s on a 100 Mbps line, it is not underperforming; it is close to the ceiling. Mistaking the two units makes a plan look eight times slower than it is.

The capital letter is the clue: lowercase b for bits, uppercase B for bytes. Speed tests usually report in Mbps, while downloaders report in MB/s, which is the root of the confusion.

The formula and a 500 MB example

For files given in gigabytes, convert first: 1 GB is 1,000 MB in decimal terms, as used by network and storage vendors, or 1,024 MB in the binary convention used by some operating systems. The two differ by about 2.4%, which only matters if you need precision.

Real links rarely hold the full figure. Protocol overhead, Wi-Fi conditions, server limits and other devices on the network all take a share. If you get 90% of the nominal speed, the same file takes 4,000 ÷ (50 × 0.9) = 89 seconds.

Check a large download by multiplying the rate you see by the time: 6.25 MB/s for 80 seconds is exactly 500 MB, the same figure from the other direction.

Time for common file sizes

Add 10–20% for real-world overhead and for the slower side of fluctuations. Large game downloads are often limited by the server, not the line, so a fast connection may not be fully used.

Streaming: bitrate to data per hour

Video streams are described by bitrate, in Mbps, which is the same unit as your connection speed. The data used per hour follows from it directly.

The margin allows for fluctuation and for other devices sharing the line. A household with two HD streams and a video call needs the sum of their rates, not just the fastest one.

Upload time is a separate sum

Uploads use the same formula with the upload speed in place of the download speed. The trap is that consumer connections are often asymmetric: a plan advertised at 100 Mbps may have far less upload speed, sometimes a tenth, unless it is a fibre plan with symmetric service.

This is why cloud backups of large photo libraries take days on slow upload lines. Check your plan's upload figure, which is listed separately from download.

The monthly ceiling of a connection

A line can only carry so much in a month, which puts fair-usage limits in context. At 100 Mbps, the line moves 12.5 MB every second.

No household uses the ceiling; the figure shows that caps, not the line speed, are the practical limit for heavy users. Check your plan's cap before scheduling large backups or game libraries, since going over may throttle the speed or cost extra.

Why the real time is longer

• Wi-Fi signal and distance from the router; a wired connection often tests higher.
• Shared bandwidth: other people streaming or syncing on the same network.
• Server-side limits, which may cap per-connection speed regardless of your plan.
• Overheads in the network protocols, typically a few percent.
• Throttling after a fair-usage threshold on some mobile and broadband plans.
• Storage speed on the receiving device, noticeable on slow SD cards or old drives.

If your measured speed is consistently far below the plan, run a speed test on a wired connection to separate the line from the Wi-Fi, and compare it with what your provider states as the typical figure.

Getting closer to the quoted speed

If a download is slower than the sum says it should be, work through the chain from the device outward. Start at the Wi-Fi: a phone in another room, behind walls, may see a fraction of the plan speed. Moving closer to the router, or switching to the 5 GHz band where available, often doubles the result. A cable connection removes the radio from the picture altogether.

Next, look at the network itself. Streaming, cloud backups and system updates on other devices all share the same line, and a single paused backup can give back a surprising amount of bandwidth. Restarting the router clears some faults, and a speed test taken on a wired laptop tells you whether the line or the Wi-Fi is the weak link.

Then consider the source. A game store or software mirror may hand you only a slice of its capacity during a busy release, and nothing at your end will improve that. Trying again at a quieter hour, or choosing a closer download region where the option exists, can help.

Lastly, check the destination. A slow memory card or a nearly full drive can limit how fast the data is written, and some devices pause downloads when the screen is off to save power. If the download time is twice what the calculation predicts, and the speed test matches the plan, the bottleneck is almost always in one of these places, not in the arithmetic.