A byte is a fixed number of bits in a group. Although the size of the group has varied over the history of the word, today a byte is universally understood to mean a collection of 8-bits. When less ambiguity is required (e.g., when defining protocols), the term octet is also sometimes used.
A Matter of Scale
When referring to storage systems (e.g., CD-ROM, DVD, hard disks, RAM, etc.) it is common to refer to the storage in terms of bytes of capacity. This generates a need to have shorthand notations for referring to large numbers of bytes. Below are the more common terms found in use, and some approximate information equivalents. Note that the abbreviations for these references all use an uppercase B to distinguish them from the equivalent notations in bits (commonly used for transmission systems).
- A byte is eight bits: the equivalent of one character (letter, number, punctuation)
- A kilobyte (kB) is approximately a thousand bytes (10^3rd or 2^10th): the equivalent of approximately one page of a standard paperback novel
- A megabyte (MB) is approximately a million bytes (10^6th or 2^20th): the equivalent of a 1,000 page novel (War & Peace?)
- A gigabyte (GB) is approximately a billion bytes (10^9th or 2^30th): a 35' shelf of paperback novels
- A terabyte (TB) is approximately a trillion bytes (10^12th or 2^40th): half of the contents of the average U.S. research library
- A petabyte (PB) is approximately a thousand, trillion bytes (10^15th or 2^50th): half the contents of all U.S. research libraries
- A exabyte (EB) is approximately a million, trillion bytes (10^18th or 2^60th): the amount of information generated worldwide in three months
- A zettabyte (ZB) is approximately a billion, trillion bytes (10^21st or 2^70th): A mind-blowing amount of information!
- A yottabyte (YB) is approximately a trillion, trillion bytes (10^24th or 2^80th): Don't even try to think about it!
In computer science, a nibble is usually four bits, or half of a byte. A word is several bytes, usually a power of two. Different computers had different word sizes (e.g., two-byte words, four-byte words, etc.). It was the common use of the four-byte word in the 1970s that led to the IP address being four bytes (32 bits) long.