Conversion from Megabits to Tebibits

Information about the Megabit (Mbit)

The megabit is a unit of digital information equal to one million bits. Its symbol is Mbit. It is the unit in which the speed of an internet connection is almost always advertised, which makes it one of the few data units that ordinary consumers encounter by name every time they choose a service.

A megabit is one million bits exactly, not 1,048,576. That distinction has practical consequences. A megabit holds 125,000 octets, which is 125 kilooctets, so a connection running at 100 megabits per second transfers about 12.5 megaoctets per second at best. A file listed as 500 megaoctets therefore takes a minimum of forty seconds, not the four the advertised number seems to promise.

That factor of eight is the reason so many people believe their connection is slower than they were sold. Nothing dishonest is happening: the industry quotes throughput in bits per second because that is what the physical layer actually carries, while file managers quote size in octets because that is how storage is organised. Both conventions are correct in their own domain, and the arithmetic between them is a division by eight.

Real throughput is lower still. Protocol headers, error correction and retransmission all consume capacity, and the usable share of a link is typically 90 to 95 per cent of its nominal rate. Wireless links lose more, because the medium is shared and interference forces retries. A connection advertised at 100 megabits per second commonly delivers 90 or so in practice, and less over a busy wireless network.

The numbers that define the eras are worth remembering. Early broadband offered 1 to 8 megabits per second, cable and fibre pushed that to 50 and 100, and gigabit services are now common in cities. High-definition video streaming needs roughly 5 megabits per second, ultra-high-definition roughly 25, and a video call between two 8, so a household's real requirement is usually far below what it buys.

In memory the megabit describes chip capacity. A 512-megabit memory chip holds 64 megaoctets, and several such chips make a module. Manufacturers count in bits because that is what the silicon holds; buyers count in octets because that is what the operating system reports.

One megabit equals 1,000,000 bits, 1,000 kilobits, 125 kilooctets, or about 0.9537 mebibits.

Information about the Tebibit (Tibit)

The tebibit is a unit of digital information equal to 1,099,511,627,776 bits, which is 1,024 gibibits or two to the fortieth power. Its symbol is Tibit. It is the binary counterpart of the terabit, and the two now differ by 10 per cent, the point at which the distinction stops being pedantic and starts being financial.

That ten per cent is the reason the IEC prefixes exist. At the kibibit the discrepancy was 2.4 per cent and could be waved away; here it is a tenth of the quantity. In a contract for storage hardware, in a service agreement that guarantees capacity, or in a specification that a supplier must meet, a tenth is the difference between compliance and breach.

The unit's real home is flash memory, where dies are now made with capacities of one tebibit and above by stacking cell layers vertically. A modern flash die may hold a tebibit or two, and a package containing eight of them holds a teraoctet of raw capacity, from which spare blocks and management overhead are subtracted before anything reaches the user.

A tebibit is 137,438,953,472 octets, or 128 gibioctets. In practical terms it is the storage of a mid-range phone, or a couple of hours of uncompressed high-resolution video. The unit describes the raw material of storage devices rather than anything a person handles directly.

Filesystems also count in binary at this scale. A volume reported as 16 tebibytes by one tool and 17.6 terabytes by another is the same volume, and the difference is entirely in the base of the arithmetic. Anyone administering storage learns to check which convention a tool uses before comparing two numbers from different sources.

Correct notation is Tibit, with the capital T and lowercase i. It appears in flash memory datasheets, in filesystem documentation and in standards, and almost nowhere else. The presence of that lowercase i is the only reliable signal that a figure is binary, which is why careful technical writing uses it even when the difference seems small.

One tebibit equals 1,099,511,627,776 bits, 1,024 gibibits, 137,438,953,472 octets, or about 1.100 terabits.