A UUID (Universally Unique Identifier) is a 128-bit identifier standardised by RFC 9562. It is designed to be globally unique without a central authority. UUIDs are written as 32 hexadecimal digits separated by hyphens in the form xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx.

Is a UUID the same as a GUID?

Yes. GUID (Globally Unique Identifier) is Microsoft's term for the same concept. The format and generation are identical — the two terms are interchangeable.

Can two UUIDs ever collide?

Theoretically yes, but the probability is negligible. A UUID v4 has 122 random bits — you would need to generate approximately 2.7 quintillion UUIDs to have a 50% chance of a single collision.

Should I use UUID v4 or v7 for database primary keys?

UUID v7 is the better choice for database primary keys because it embeds a millisecond Unix timestamp, making UUIDs time-sortable. This avoids the random page splits and index fragmentation caused by v4 UUIDs, especially in B-tree indexes.

← Learn

What Is a UUID?

Updated 8 May 2026

A plain-English guide to UUIDs — what they are, how they are structured, the difference between v4 and v7, and when to use a UUID as a database primary key.

What a UUID is

A UUID (Universally Unique Identifier) is a 128-bit number used to identify things — database rows, API resources, session tokens, files, and anything else that needs a globally unique label. The defining feature is that UUIDs can be generated on any machine at any time without checking a central registry, yet collisions are so statistically rare they are treated as impossible in practice.

UUID format — 32 hex digits, 4 hyphens, 5 groups

550e8400-e29b-41d4-a716-446655440000

xxxxxxxx — 8 hex = time_low (v7) or random (v4)

xxxx — 4 hex = time_mid

4xxx — version digit + 3 hex

yxxx — variant bits + 3 hex

xxxxxxxxxxxx — 12 hex = random node

UUID vs GUID — the same thing

Microsoft calls them GUIDs (Globally Unique Identifiers). The format and generation algorithm are identical. You will see both terms depending on the technology stack — UUID in Linux/web/database contexts, GUID in Windows/.NET contexts.

UUID versions

Version	How it is generated	Use today?
v1	MAC address + timestamp	Avoid — leaks hardware identity
v3	MD5 hash of namespace + name	Legacy only
v4	Fully random (122 random bits)	Yes — widely supported default
v5	SHA-1 hash of namespace + name	When deterministic IDs needed
v7	Unix timestamp (ms) + random	Yes — preferred for databases

Why UUID v7 is better for databases

UUID v4 is purely random. When used as a database primary key in a B-tree index, new rows are inserted at random positions — this causes page splits and index fragmentation, which degrades write performance at scale.

UUID v7 embeds a millisecond Unix timestamp in the first 48 bits. This means new UUIDs are always greater than earlier ones, so inserts go to the end of the index — the same pattern as an auto-incrementing integer. You get global uniqueness without the performance penalty.

Common uses

Database primary keys — UUID v7 is ideal; v4 works but fragments indexes
API resource identifiers — expose UUIDs rather than sequential integers to avoid enumeration attacks
Session tokens and correlation IDs — easy to generate, globally unique across distributed systems
Idempotency keys — generate a UUID client-side so retried requests are deduplicated server-side
File names — avoid collisions when uploading files from multiple sources

Related guides

UUID v4 vs v7 — Which Should You Use? →

Generate UUID v4 and v7 online

Bulk generate, copy all, toggle uppercase or braces — instantly in the browser

Open UUID Generator →