Centralized Unique ID

⸺ by Charles Iliya Krempeaux

⸺ up

⸺ previous

⸺ next

There are two main strategies for creating a unique identifier (unique ID):

distributed, and
• centralized.

There are also some strategies that are some form of hybrid between the two.

Some examples of unique-IDs created using a centralized registry are:

IETF RFC numbers,
• URL shortener IDs (under a specific Internet domain name),
• the Internet domain name system (if you include time),
• phone numbers (if you include time).

They can sometimes work.

It is possible for them to be more space-efficient (if their designers choose to design them to be space-efficient).

But they also have major problems if you are not in total control over the registry. For example — your unique ID can be taken from you, or censored, etc.

(This has happened with Internet domain names a number of times!)

And in many contexts this major problem with centralized unique IDs makes them completely unacceptable to be used.

UUID

UUIDs, in general, are NOT centralized.

UUIDs, in general, are either distributed, or a hybrid.

One of the design goals of UUID was for them to be able to be created in a distributed manner.

And with UUIDs of type version=4 and variant=10 it is possible to create UUIDs in a completely distributed manner.