What do you think of Universally Unique Lexicographically Sortable Identifier (ULID)?

Using FoundationDB
1

According to https://github.com/ulid/spec:

UUID can be suboptimal for many use-cases because:

  • It isn’t the most character efficient way of encoding 128 bits of randomness
  • UUID v1/v2 is impractical in many environments, as it requires access to a unique, stable MAC address
  • UUID v3/v5 requires a unique seed and produces randomly distributed IDs, which can cause fragmentation in many data structures
  • UUID v4 provides no other information than randomness which can cause fragmentation in many data structures

Instead, herein is proposed ULID:

  • 128-bit compatibility with UUID
  • 1.21e+24 unique ULIDs per millisecond
  • Lexicographically sortable!
  • Canonically encoded as a 26 character string, as opposed to the 36 character UUID
  • Uses Crockford’s base32 for better efficiency and readability (5 bits per character)
  • Case insensitive
  • No special characters (URL safe)
  • Monotonic sort order (correctly detects and handles the same millisecond)

Do you use it? Do you recommend it?

1 Like
2

Ah. I think the “:” somehow attached itself to the URL which is causing it to 404. Here’s a sanitized one: https://github.com/ulid/spec

3

I guess one thing I’ll point out: the “UUID” type used by the Tuple bindings is, I think, compatible with this spec (at least when the data are serialized). It encodes a 128 bit UUID as one (1) byte of type code followed by sixteen (16) bytes of data, and it isn’t particularly opinionated about what the data look like. This will then sort your UUIDs according to their unsigned lexicographic byte order, which I think is sufficient for this format. You might need some extra scaffolding to transform between UUIDs from the ULID spec and UUIDs that fell off a truck (or wherever), but I think it would do most of the low-level things, etc.

Uh, assuming that the spec is using unsigned 48-bit integers for the timestamp.

1 Like
4

I think you’d need to be careful that ULID has a timestamp as a leading part, so inserting a series of ULIDs into a database would cause a rotating write hotspot, whereas inserting UUID v4/v5 would distribute over the keyspace well.

2 Likes