What Is a UUID and When to Use It in Your Development Projects

What exactly is a UUID

UUID stands for Universally Unique Identifier — a 128-bit identifier written as 32 hexadecimal characters, grouped in the standard 8-4-4-4-12 format (for example 550e8400-e29b-41d4-a716-446655440000). It is defined in the RFC 4122 standard, which specifies several "versions" depending on how the value is generated.

The most common version today is 4 (UUID v4), where practically all the bits are randomly generated. Other versions (1, 3, 5, 7) use combinations of timestamps, network addresses, or name hashes — each with different purposes, but v4 is the de facto standard when you just need "a unique identifier, right now".

Why not simply use 1, 2, 3...

An incremental (autoincrement) ID has a structural limitation: it requires central coordination. Only one database can decide "the next number is 47" without risking duplicates. This works perfectly in a system with a single database, but becomes a problem in distributed systems — microservices, offline-first applications, synchronization between devices — where multiple sources need to generate IDs independently without talking to each other.

UUIDs solve this by design: any system, anywhere, without coordinating with anyone else, can generate a UUID with a collision probability so low it is statistically irrelevant (1 in 5.3×10³⁶ for UUID v4). Also, an incremental ID reveals information (how many records exist, in what order they were created), while a UUID does not leak that.

Drawbacks to keep in mind

UUIDs are not free: they take more storage (16 bytes vs 4-8 for an integer), and as a primary key in some relational databases they can degrade B-tree index performance because the values are not sequential (this is mitigated with UUID v7, which includes a sortable timestamp at the start). They are also harder to read or say out loud compared with "order #47".

The practical decision: use UUIDs when you need decentralized generation, non-guessable IDs (additional security through obscurity), or you will merge data from multiple sources. Use incremental IDs when you have a single source of truth and the performance of sequential indexes matters more than decentralization.

Frequently asked questions

Are all UUIDs equally random?

No — only version 4 is mostly random. Other versions incorporate timestamps or hashes for specific ordering or determinism purposes.

Can I use a UUID as a primary key in any database?

Yes, technically, but in relational databases with B-tree indexes it can affect large-scale insert performance. UUID v7 (time-sortable) mitigates this problem.

Can a UUID repeat?

The probability is so low that it is considered practically impossible in any real application — you would need to generate billions per second for decades to have a meaningful chance of a collision.

How do I generate a UUID in JavaScript?

With crypto.randomUUID(), available natively in all modern browsers and in Node.js, with no external libraries needed.