| Zebibits (Zibit) | Exbibits (Eibit) |
|---|---|
| 1 Zebibit | 1024 Eibit |
| 2 Zebibits | 2048 Eibit |
| 3 Zebibits | 3072 Eibit |
| 4 Zebibits | 4096 Eibit |
| 5 Zebibits | 5120 Eibit |
| 10 Zebibits | 10240 Eibit |
| 20 Zebibits | 20480 Eibit |
| 25 Zebibits | 25600 Eibit |
| 50 Zebibits | 51200 Eibit |
| 100 Zebibits | 102400 Eibit |
| Reference | Zebibits (Zibit) | Exbibits (Eibit) |
|---|---|---|
| A plain text message (160 characters) | 1.0842 × 10-18 Zibit | 1.11022 × 10-15 Eibit |
| A three-minute MP3 | 2.03288 × 10-14 Zibit | 2.08167 × 10-11 Eibit |
| A smartphone photo | 2.71051 × 10-14 Zibit | 2.77556 × 10-11 Eibit |
| A high-definition film | 2.71051 × 10-11 Zibit | 0.0000000277556 Eibit |
| A dual-layer Blu-ray disc | 3.38813 × 10-10 Zibit | 0.000000346945 Eibit |
The zebibit is a unit of digital information equal to two to the seventieth power bits, which is 1,024 exbibits. Its symbol is Zibit. It is the binary counterpart of the zettabit, and the two now differ by 18.1 per cent — nearly a fifth.
That divergence is the reason the binary prefixes were needed at all. At the kibibit the two conventions differed by 2.4 per cent, a rounding error. Here the difference is large enough that a document using the wrong one is simply reporting a different quantity, and no amount of context can repair the ambiguity after the fact.
A zebibit is 147,573,952,589,676,412,928 octets, or 128 exbioctets. Nothing of this size exists. The total data held by humanity is somewhere in the low hundreds of zettaoctets, which is under a zebioctet, so the world's entire information stock does not yet reach one unit at this step of the binary ladder.
The unit exists because the IEC series was defined completely rather than as far as anyone then needed. That is the same principle the metric system follows: every prefix applies to every unit, whether or not the combination has yet been used. A system with holes in it requires a table of exceptions, and a system without holes requires only the rule.
Where the zebibit could genuinely appear is in address space arithmetic. Two to the seventieth is not a natural pointer width, but multiples and fractions of powers of two run through every discussion of addressing, and a scheme that reserved seventy bits for something would naturally be described in these terms. Such schemes are proposed occasionally and none has yet been needed.
Reading the symbol is the practical skill. Zibit is binary, Zbit is decimal, and at eighteen per cent apart the two are not interchangeable in any document where the number matters. Where a source writes ZB with no explanation, there is no way to know which was meant, and the honest response is to treat the figure as uncertain to a fifth.
One zebibit equals 1,024 exbibits, 147,573,952,589,676,412,928 octets, or about 1.181 zettabits.
The exbibit is a unit of digital information equal to two to the sixtieth power bits, which is 1,024 pebibits. Its symbol is Eibit. It is the binary counterpart of the exabit, and the two now differ by 15.3 per cent.
Two to the sixtieth is a number with a particular importance in computing, because sixty is close to the sixty-four bits of a modern processor address. A 64-bit machine can address sixteen exbioctets of memory, and that figure — 16 EiB — is the theoretical ceiling of the entire architecture. It appears in processor manuals, in operating system documentation and in the specification of every 64-bit filesystem.
No machine comes close to using that range. Current processors implement only forty-eight or fifty-seven of those sixty-four address bits, because wiring the full width would cost silicon for an address space nobody can fill. The unused bits are reserved, and widening the implementation is a straightforward matter whenever memory sizes make it worthwhile, which is the point of having chosen sixty-four in the first place.
An exbibit is 144,115,188,075,855,872 octets, or 128 pebioctets. That is more storage than any single organisation holds, and comparable to the combined annual output of a large part of the storage industry. As with all the larger binary units, it describes limits and capacities in specifications rather than anything that has been built.
The unit is also where filesystem designers set their maximum sizes. Filesystems built around 64-bit block pointers naturally have limits at exact powers of two, and several widely deployed ones specify maximum volume sizes in exbioctets. Those numbers are not aspirations; they are the arithmetic consequence of the pointer width, and they will hold until the architecture changes.
Reading the symbol correctly matters here. Eibit is the exbibit; Ebit is the exabit; EB and EiB are the octet forms of each. In a document where a fifteen per cent difference is significant — and at this scale it always is — the presence or absence of the lowercase i carries the entire meaning.
One exbibit equals 1,024 pebibits, 144,115,188,075,855,872 octets, or about 1.153 exabits.