A set of prefixes for binary powers designed to be used in data processing and data transmission contexts. They were suggested in 1998 by the International Electrotechnical Commission (IEC) as a way of resolving the ambiguity in use of kilo-, mega-, giga-, etc., in computing. In scientific usage, these prefixes indicate 103, 106, 109, etc. (see SI units). In computing, it became common to use the prefix ‘kilo-’ to mean 210, so one kilobit was 1024 bits (not 1000 bits). This was extended to larger prefixes, so ‘mega-’ in computing is taken to be 220 (1 048 576) rather than 106 (1 000 000). However, there is a variation in usage depending on the context. In discussing memory capacities megabyte generally means 220 bytes, but in disk storage (and data transmission) megabyte is often taken to mean 106 bytes. (In some contexts, as in the capacity of a floppy disk, it has even been quoted as 1 024 000 bytes, i.e. 1000 times a (binary) kilobyte.) The IEC attempted to resolve this confusion by introducing binary prefixes, modelled on the normal decimal prefixes, as follows:
These names are contractions of ‘kilobinary’, ‘megabinary’, etc., but are pronounced so that the second syllable rhymes with ‘bee’. Using these prefixes, one gebibyte would be 1 073 741 824 bytes and one gigabyte would (unambiguously) be 1 000 000 000 bytes.