ULID

ULID (Universally Unique Lexicographically Sortable Identifier) is an alternative to UUIDs that provides better sortability and readability.

What is ULID?

ULID is a 128-bit identifier like UUID, but with several advantages:

• Lexicographically sortable: Natural sort order matches creation time
• Compact: 26 characters vs 36 for UUID (no hyphens)
• URL-safe: Uses Crockford’s Base32 alphabet
• Case-insensitive: Reduces errors in manual entry
• Timestamp-based: First 48 bits encode Unix timestamp (millisecond precision)
• Random component: Last 80 bits ensure uniqueness

Example ULID: 01ARZ3NDEKTSV4RRFFQ69G5FAV

Configuring ULID Generator

Enable ULID generation in your application:

use Cline\StronglyTypedId\Enums\GeneratorType;
use Cline\StronglyTypedId\Facades\IdGenerator;

IdGenerator::setGenerator(GeneratorType::Ulid);

Now all ID generation will use ULIDs:

$userId = UserId::generate();
// e.g., "01ARZ3NDEKTSV4RRFFQ69G5FAV"

Creating IDs from ULID Strings

ULIDs are compatible with the standard fromString() method:

$userId = UserId::fromString('01ARZ3NDEKTSV4RRFFQ69G5FAV');

ULID Format Structure

01ARZ3NDEKTSV4RRFFQ69G5FAV
|----------|--------------|
 Timestamp      Randomness
  10 chars       16 chars

Timestamp Component (48 bits):

• First 10 characters
• Millisecond precision Unix timestamp
• Provides natural chronological ordering

Random Component (80 bits):

• Last 16 characters
• Cryptographically strong randomness
• Ensures uniqueness

Chronological Ordering

ULIDs sort naturally by creation time:

$id1 = UserId::generate(); // 01ARZ3NDEKTSV4RRFFQ69G5FAV
sleep(1);
$id2 = UserId::generate(); // 01ARZ3NDEKTSV4RRFFQ69G5FB0

// String comparison reflects chronological order
$id1->toString() < $id2->toString(); // true

This makes ULIDs excellent for database indexing and time-based queries.

ULID vs UUID Comparison

Advantages of ULID

Feature	ULID	UUID
Length	26 chars	36 chars
Sortable	Yes (chronological)	Only v6/v7
URL-safe	Yes (Base32)	No (hyphens)
Case-sensitive	No	Yes
Index performance	Excellent	Good (v7) / Poor (v4)

Advantages of UUID

• Wider adoption and tooling support
• Native database types
• Multiple specialized versions (v1-v8)
• Name-based deterministic generation (v3/v5)

Database Storage

String Storage

Store ULIDs as CHAR(26) or VARCHAR(26):

Schema::create('users', function (Blueprint $table) {
    $table->char('id', 26)->primary();
});

Binary Storage (More Efficient)

Convert ULIDs to binary for storage efficiency:

use Ulid\Ulid;

// Convert to binary (16 bytes)
$binary = Ulid::fromString($userId->toString())->toBytes();

// Store in database
DB::table('users')->insert(['id' => $binary]);

Binary storage saves 10 bytes per record (26 chars vs 16 bytes).

Monotonicity Guarantee

ULIDs generated within the same millisecond maintain monotonic ordering:

$ids = [];
for ($i = 0; $i < 100; $i++) {
    $ids[] = UserId::generate();
}

// All IDs maintain strict increasing order
for ($i = 1; $i < count($ids); $i++) {
    assert($ids[$i-1]->toString() < $ids[$i]->toString());
}

Timestamp Extraction

Extract timestamps from ULIDs when needed:

use Ulid\Ulid;

$userId = UserId::generate();
$ulid = Ulid::fromString($userId->toString());

$timestamp = $ulid->toDateTime();
$milliseconds = $ulid->getTimestamp();

Use Cases

Choose ULID when:

• High-volume applications needing indexing performance
• User-facing IDs (more readable, URL-safe)
• Time-series data with chronological ordering
• Distributed systems needing ordering across nodes

Choose UUID when:

• Legacy system compatibility
• Native UUID database column types needed
• Deterministic generation required (v3/v5)
• RFC 4122 compliance required