Upgrade the NewV7 Implementation to Align with PostgreSQL

[sergeyprokhorenko]

The current implementation should be upgraded to match the most advanced official implementation available for PostgreSQL, which offers superior functionality.

See the article: Postgres UUIDv7 + Per-Backend Monotonicity

[rowanseymour]

Is it superior functionality or different? My understanding is that the current implementation in this library (Method 1) guarantees sortability within a single process, at the expense of time accuracy since 1) time component is just millisecond accuracy and 2) the time value is bumped into the future when more than 4096 values are generated in the same millisecond.. which I guess is allowed in the RFC?

Method 3 that PostgreSQL moved to uses rand_a for increased time accuracy but now the values from a single process are not guaranteed sortable because there's no guarantee that subsequent calls use a time value greater than the previous... or is that not the case?

[sergeyprokhorenko]

@rowanseymour

No, it offers not just different but superior functionality. The PostgreSQL algorithm provides additional guarantees compared to RFC 9562. In fact, it combines Method 3 with Method 1. The entire timestamp can also function as a counter when more than about 4 identifiers per microsecond are generated. Therefore, the monotonicity of generated UUIDs is ensured within the same backend (a single process).

A mutex could have ensured monotonicity across all backends, but it was too slow, so it was abandoned.

Monotonically increasing identifiers are generated even if the system clock jumps backward, if access to the system clock is unavailable, or if UUIDs are generated at a very high frequency, due to the internal timestamp functioning as a counter to maintain order.

The uuidv7() function can accept an optional offset parameter of type interval which is added to the internal timestamp. If the offset parameter results in a timestamp overflow or a negative timestamp, an adjusted timestamp value is automatically used. The timestamp behaves like a ring buffer: when the maximum value is exceeded, it wraps around to the minimum value. Similarly, if the absolute value of the negative offset exceeds the time elapsed since 00:00:00 UTC on January 1, 1970, the timestamp wraps around to the maximum value.

See the article: Postgres UUIDv7 + Per-Backend Monotonicity

[rowanseymour]

@sergeyprokhorenko do you know of any implementations in go that match the PostgresSQL algorithm?

[sergeyprokhorenko]

@rowanseymour

There are no such implementations yet, but there is a discussion.

Note that this algorithm requires finding access to a high-precision system clock. Andrey Borodin, the developer of uuidv7() for PostgreSQL, has done this work for all platforms.

What May Surprise You About UUIDv7

[rowanseymour]

@sergeyprokhorenko golang/go#62026