UUID

A UUID (Universally Unique Identifier) ​​is a standardized identifier designed to be unique and universal. It is defined by RFC 4122 and adopted by the ISO/IEC 9834-8:2008 standard.

A UUID consists of 128 bits of data and is generally represented in hexadecimal format using the canonical format xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx.

Depending on the version used, the uniqueness of a UUID can be guaranteed deterministically (by design) or probabilistically (by random generation).

A UUID can be generated according to different versions:

Version 4 is generated from pseudo-random numbers. Of the 128 bits, 122 bits are random, the others being reserved for the version and variant. Uniqueness is therefore probabilistic but extremely high.

Version 3 is generated from an MD5 hash applied to the concatenation of two elements: a UUID named namespace and a string representing the element to be encoded.

Version 5 follows the same principle as version 3 but uses the SHA-1 hashing algorithm instead of MD5.

For versions 3 and 5, the generation is deterministic: for the same namespace and the same name (string of characters), the UUID produced is always identical.

A version 4 UUID is based on random bits and contains no usable semantic information. Therefore, there is nothing to decode besides its version and variant.

A version 3 or 5 UUID is generated by the cryptographic hashing of a string of characters combined with a namespace. Since MD5 and SHA-1 are one-way hash functions, it is not possible to recover the original string from the UUID.

However, regardless of the UUID type, certain bits always allow the version and variant to be identified.

UUIDs are generally categorized into different versions, ranging from version 1 to version 5.

Version 1: contains a timestamp and MAC address

Version 2: a variant of V1 for DCE (Distributed Computing Environment) which includes POSIX UIDs

Version 3: MD5 hash

Version 4: randomly generated

Version 5: SHA-1 hash

UUIDs also have a variant number (3 bits) assigned to them:

0xx: Reserved for NCS (Network Computing System) for backward compatibility

10x: Default variant, see RFC 4122

110: Reserved for Microsoft for backward compatibility

111: Reserved for future definition

A UUID string represents 128 bits encoded as 32 hexadecimal characters, typically divided into 5 groups of hexadecimal digits separated by hyphens -, according to the 8-4-4-4-12 pattern: xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx

A string strictly adhering to this format can be recognized syntactically as a UUID, but this does not guarantee that it was generated correctly.

Some characters in a UUID string explicitly encode the version and variant. In the canonical representation xxxxxxxx-xxxx-Yxxx-Zxxx-xxxxxxxxxxxx, each hexadecimal character corresponds to half a byte (4 bits).

The hexadecimal character Y corresponds to the 13th half-byte of the string. It directly encodes the UUID version.

The hexadecimal character Z corresponds to the 17th half-byte of the string. It encodes the UUID variant. The 3 most significant bits of this half-byte determine the variant (see above).

A version 4 UUID has 122 random bits, which corresponds to 2^122 possible values (on the order of 10^36).

Even generating several billion UUIDs per second for centuries, the probability of a collision remains negligible.