A network connection advertised at 100 megabits per second and a download window showing 12 megabytes per second may be describing almost the same rate. The difference is the unit. One byte contains eight bits, so 100 megabits per second corresponds to 12.5 megabytes per second before allowing for protocol overhead and other limits.
Capitalization can carry the distinction: “b” commonly means bit and “B” means byte. Not every interface labels units consistently, so the full term is more reliable than guessing from a large number alone.
Convert before comparing
For decimal units, a megabit is one million bits and a megabyte is one million bytes. Divide a rate in megabits per second by eight to express the same rate in megabytes per second. Multiply a rate in megabytes per second by eight to express it in megabits per second.
| Displayed rate | Equivalent decimal rate |
|---|---|
| 40 Mb/s | 5 MB/s |
| 80 Mb/s | 10 MB/s |
| 200 Mb/s | 25 MB/s |
| 1 MB/s | 8 Mb/s |
These are unit conversions, not speed-test results. Nothing in the table says a particular service will deliver its advertised maximum continuously. It simply puts the labels on the same scale so a meaningful comparison can begin.
A file-size estimate that shows its work
Suppose a file is exactly 600 MB and the usable transfer rate remains exactly 10 MB/s. The ideal transfer time is 600 divided by 10, or 60 seconds. If the rate is stated as 80 Mb/s, convert it to 10 MB/s first and the same estimate follows.
Dividing 600 MB by 80 Mb/s without converting would produce an incorrect 7.5-second result. The bytes and bits do not cancel as written. Carrying the unit through the arithmetic exposes the mistake before it becomes an expectation about the connection.
The calculation assumes a constant usable rate. Real transfers may pause, ramp up, share a connection, or wait for the remote server. An app may also be verifying or unpacking a download after the network transfer ends. Those stages take time without necessarily showing continuous incoming data.
Why a gigabyte can appear to shrink
Decimal prefixes and binary prefixes are a second, separate issue. One gigabyte, written GB, is one billion bytes. One gibibyte, written GiB, is 1,073,741,824 bytes. The underlying count of bytes can be unchanged while its numerical expression differs between those units.
A hypothetical device containing exactly 1,000,000,000 bytes contains about 0.931 GiB. That conversion alone does not prove that capacity vanished. Usable storage can also be reduced by formatting, system files, reserved space, and other allocations, but those are additional effects that should not be confused with the unit conversion.
NIST documents the binary prefixes precisely because using “kilo” or “mega” for both decimal and binary multiples creates ambiguity. When a precise comparison matters, obtain the actual byte count and the unit convention used by each display.
Rates, quotas, and duration are different fields
A service allowance of 100 GB describes an amount of data. A rate of 100 Mb/s describes how quickly data can move. A monthly allowance does not tell you the connection speed, and a speed label does not tell you whether a quota applies.
Likewise, download and upload rates describe opposite directions. A cloud backup depends heavily on upload performance, while receiving a large file depends on download performance. A single headline number may describe only one direction.
For a clear comparison, record four items: file size with its unit, rate with its unit, transfer direction, and whether the rate is an advertised maximum or an observed usable rate. Then do the arithmetic. That short discipline explains many apparent discrepancies without assuming either that a service is broken or that a favorable speed test predicts every future transfer.
Sources
- NIST: Binary prefixes
One byte is eight bits; decimal and binary prefixes represent different multiples.
- NIST: Decimal prefixes
SI prefixes represent powers of ten rather than powers of two.