Understanding ULID: A Lexicographically Sortable Unique Identifier and Its Python Implementation

ULID (Universally Unique Lexicographically Sortable Identifier) is presented as an alternative to UUID, providing a 128‑bit identifier that combines a millisecond‑precision timestamp with a large random component, resulting in lexicographically sortable, URL‑safe IDs with a negligible risk of collision.

Unlike UUID’s five versions—where version 1 requires a MAC address, version 2 replaces part of the timestamp with UID/GID, versions 3 and 5 rely on hash functions, and version 4 is purely random—ULID merges time and randomness, offering better ordering and readability.

Key ULID characteristics:

128‑bit compatibility with UUID

1.21×10²⁴ unique IDs per millisecond

Lexicographic (dictionary) ordering

Encoded as a 26‑character Crockford Base32 string (no I, L, O, U)

Case‑insensitive and URL‑safe

Monotonic ordering within the same millisecond

ULID specification: The identifier consists of a 48‑bit timestamp (UNIX time in milliseconds) and an 80‑bit random component. The binary layout is encoded in network byte order across 16 bytes.

01AN4Z07BY      79KA1307SR9X4MV3
|----------|    |----------------|
Timestamp          Randomness
10 chars           16 chars
48 bits            80 bits

Components:

Timestamp

48‑bit integer representing UNIX time in milliseconds

Valid until the year 10889

Randomness

80‑bit random number, preferably generated with a cryptographically secure source

Sorting: Characters are ordered lexicographically; the leftmost character is the most significant. Within the same millisecond, strict ordering cannot be guaranteed.

Encoding: Uses Crockford’s Base32 alphabet (0123456789ABCDEFGHJKMNPQRSTVWXYZ) to improve efficiency and readability.

0123456789ABCDEFGHJKMNPQRSTVWXYZ

Binary layout and byte order: The 16‑byte representation is big‑endian (network byte order).

+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|          32_bit_uint_time_high          |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|   16_bit_uint_time_low   |   16_bit_uint_random   |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|          32_bit_uint_random            |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|          32_bit_uint_random            |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

Typical use cases:

Replace auto‑increment primary keys in databases without DB involvement

Distributed systems needing globally unique, time‑ordered IDs

Database sharding/partitioning based on embedded timestamps

Sorting records by ULID instead of a separate timestamp column when millisecond precision suffices

Python usage (ulid‑py library):

pip install ulid-py

Generate a new ULID:

>> import ulid
>>> ulid.new()

Convert an existing UUID to ULID:

>> import ulid, uuid
>>> value = uuid.uuid4()
>>> ulid.from_uuid(value)

Create a ULID from a specific timestamp:

>> import datetime, ulid
>>> ulid.from_timestamp(datetime.datetime(1999, 1, 1))

Generate a ULID from custom randomness:

>> import os, ulid
>>> randomness = os.urandom(10)
>>> ulid.from_randomness(randomness)

Access ULID components:

>> u = ulid.new()
>>> u.timestamp()
>>> u.randomness()

For more information, see the project repository at https://github.com/ahawker/ulid .