What are these different types of sizes?
In particular, what is a kebibyte, mebibyte, etc? It does sound made up. But these prefixes are used to resolve historical differences between usage of the "kilo" prefix as used in other weights and measures (such as kilogram), which means 1000, and computing (especially with rerference to computer memory) which generally uses 1024. This was parodied by xkcd 394, which has a humorous table of different so-called kilobyte standards and included a table which is also included below (but don't use it as a reference :))
| Symbol | Name | Size | Notes |
|---|---|---|---|
| kB | Kilobyte | 1024 or 1000 bytes | 1000 bytes during leap years, 1024 otherwise |
| KB | Kelly-Bootle Standard Unit | 1012 bytes | Compromise betweeen 1000 and 1024 bytes |
| KiB | Imaginary Kilobyte | 1024√-1 bytes | Used in Quantum Computing |
| kB | Intel Kilobyte | 1023.937528 bytes | Calculated on Pentium FPU |
| Kb | Drivemaker's Kilobyte | Currently 908 bytes | Shrinks by 4 bytes each year for marketing reasons |
| KBa | Baker's Kilobyte | 1152 bytes | 9 bits to the byte since you're such a good customer |
Base 10 (1000 bytes = 1 kilobyte)
This definition is generally used by drive manufacturers and is the standard SI unit for a kilobyte. It's used in data transfer rates and some other things. The advantages of it is that it's a SI unit, so it's likely to be understood. Some disadvantages include that it is inconsistent with how computer memory is reported (using base 2 instead) and it is inconsistent with some popular operating systems. This page uses the SI prefixes, disambiguating by using the "kibi" suffixes.
Base 2 (1024 bytes = 1 kilobyte)
Another popular contender is to use base 2 (i.e. 1024 bytes in 1 kilobyte). This is commonly used by memory manufactuers, CPU cach sizs, etc (due to the heavy use of binary addressing), and some telecommunications companies, and probably most importantly, Microsoft Windows. Because this page uses the SI prefixes, we disambiguate with the kibi suffix. However, as xkcd has again pointed out...
I would take 'kibibyte' more seriously if it didn't sound so much like 'Kibbles N Bits'.
In addition, many older computers report using powers of 2. Many tools on Linux, such as top, report using the "kibi" suffix.
How big is a byte?
8 bytes, generally. This is de facto standard, but some computers, particularly older computers,
used a different byte size. Many 6-bit of 9-bit computers were used in the 1960s to store characters and byte
sizes could be anywhere from 2 to 3 to 4 to 8 to 10. The term octet refers unambiguously to a grouping of
8 bits, and can be seen today in some media types such as application/octet-stream.