What is UUID?

UUID (Universally Unique Identifier) is a 128-bit globally unique identifier, formatted as 32 hexadecimal characters in 5 groups: xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx. UUIDs enable distributed systems to generate unique IDs without central coordination.

UUID Version Comparison

UUID v4 Details

UUID v4 has 2¹²² possible values, making collision probability negligible:

• 1 trillion IDs → ~10⁻¹⁸ collision probability
• 1 billion IDs/second → 10 billion years for collision

UUID v4 Limitations

• Not sortable: Poor for DB primary keys
• Index performance: Random inserts cause B+ tree page splits
• Storage: 128-bit / 36-char string, larger than auto-increment

What is ULID?

ULID (Universally Unique Lexicographically Sortable Identifier) is a sortable globally unique identifier designed to solve UUID v4's ordering problem.

ULID Features

• 128-bit: 26-char Crockford Base32 encoding
• Sortable: Lexicographically ordered by timestamp
• Millisecond precision: 48-bit timestamp
• Case-insensitive: No special characters
• URL-safe: No special characters needed

ULID Structure

01AN4Z07BY      79KA1307SR9X4MV3
│               │
│               └── Random (80 bit)
│
└── Timestamp (48 bit, ms Unix time)

UUID v7 (RFC 9562)

Standardized in May 2024, the best UUID v4 alternative:

• First 48 bits: Unix millisecond timestamp
• Remaining 74 bits: Random
• Time-sortable for database indexing
• No coordination needed for distributed generation

DB Primary Key Recommendations

Security Notes

• Use cryptographically secure RNG (CSPRNG) for UUID v4
• Avoid Math.random() for UUID generation
• Don't use UUIDs for security tokens