| Zettaoctets (Zo) | Mebioctets (Mio) |
|---|---|
| 1 Zettaoctet | 953674316406250 Mio |
| 2 Zettaoctets | 1.90734863281 × 1015 Mio |
| 3 Zettaoctets | 2.86102294922 × 1015 Mio |
| 4 Zettaoctets | 3.81469726562 × 1015 Mio |
| 5 Zettaoctets | 4.76837158203 × 1015 Mio |
| 10 Zettaoctets | 9.53674316406 × 1015 Mio |
| 20 Zettaoctets | 1.90734863281 × 1016 Mio |
| 25 Zettaoctets | 2.38418579102 × 1016 Mio |
| 50 Zettaoctets | 4.76837158203 × 1016 Mio |
| 100 Zettaoctets | 9.53674316406 × 1016 Mio |
| Reference | Zettaoctets (Zo) | Mebioctets (Mio) |
|---|---|---|
| A plain text message (160 characters) | 1.6 × 10-19 Zo | 0.000152588 Mio |
| A three-minute MP3 | 3 × 10-15 Zo | 2.86102 Mio |
| A smartphone photo | 4 × 10-15 Zo | 3.8147 Mio |
| A high-definition film | 4 × 10-12 Zo | 3814.7 Mio |
| A dual-layer Blu-ray disc | 5 × 10-11 Zo | 47683.7 Mio |
The zettaoctet is a unit of digital information equal to a thousand exaoctets, or a sextillion octets. Its symbol is Zo. It is the unit in which the total quantity of data in the world is now described, and the only data unit that has given its name to an era.
That name came from the phrase the zettabyte era, used from around 2016, when annual global internet traffic first exceeded one zettaoctet. Estimates of the global datasphere — every file stored anywhere, plus everything created and immediately discarded — passed a hundred zettaoctets a few years later and continue to rise by tens of zettaoctets a year.
Understanding what these figures include matters more than the numbers themselves. They count copies. A single photograph on a phone is also on the phone's backup, on two servers in a cloud provider's data centre, in a caching layer, and possibly in a machine learning training set. Counting stored octets rather than distinct works inflates the total by a large factor that nobody has ever measured precisely.
The composition of the total is dominated by things nobody looks at. Surveillance and security video, most of it overwritten within weeks, is the single largest category. Sensor and telemetry data from industrial equipment, vehicles and buildings is second. Human-authored material of any kind is a minority, and text is a rounding error within that minority.
Manufacturing sets a hard limit on the total. All the world's storage factories together produce a few hundred exaoctets of new capacity per year, which is a fraction of a zettaoctet, so most data created is never stored at all. It is processed as it arrives and thrown away, and this is a deliberate design choice rather than a failure.
For a sense of the scale, a zettaoctet would be about eight thousand million large consumer hard drives, or roughly one for every person on Earth. Stacked flat, that many drives would form a column reaching well beyond the Moon. The unit describes a quantity that only exists as a sum across a whole planet's industry.
One zettaoctet equals 1,000 exaoctets, 1,000,000,000,000 gigaoctets, 8 zettabits, or about 0.8470 zebioctets.
The mebioctet is a unit of digital information equal to 1,048,576 octets, which is 1,024 kibioctets. Its symbol is Mio. It is the binary counterpart of the megaoctet, and the two differ by 4.9 per cent — the point at which a careful writer starts distinguishing them.
The unit's most visible home is the processor cache. The fast memory built into a processor to hold recently used data is sized in binary, and modern chips carry caches measured in mebioctets: a few hundred kibioctets at the second level and tens of mebioctets at the third. Those figures are exact powers of two because the cache is addressed by dividing an address into fixed bit fields.
Block sizes and buffers follow the same logic. Filesystems allocate in blocks, database engines read in pages, compression tools work in windows, and all of these are powers of two, most often a few mebioctets. A tool that offers a buffer of 16 megaoctets almost always means 16 mebioctets, because the underlying allocation is a shift rather than a multiplication.
The distinction has become visible to ordinary users through the discrepancy between what a download claims and what a file manager reports. A file described as 100 megaoctets on a website is often 100 mebioctets on disc, or the reverse, and the resulting 4.9 per cent difference is enough to make a progress bar look wrong without anything actually being wrong.
For scale, a mebioctet holds a million characters of unaccented text — roughly a long novel, or five hundred pages. It is also a single photograph from a modest camera, or eight seconds of high-definition video. The same unit therefore describes both a very large amount of text and a very small amount of video, which says a good deal about the relative cost of representing the two.
The IEC notation Mio is used in technical documentation, in the Linux kernel, in filesystem tools and in standards. It remains rare in consumer software, where MB is written for both quantities. When a figure matters, the safest habit is to compute in octets and convert once at the end.
One mebioctet equals 1,048,576 octets, 1,024 kibioctets, 8 mebibits, or about 1.049 megaoctets.