Hacker News Reader

Comment by everfrustrated | Hacker News Reader

I'm scratching my head trying to think why AWS would have worked on this? I can't think of it being used in any of their products.

RDS uses block replication. Aurora uses it's own SAN replication layer.

DMS maybe?

View on HN

Apparently they've offered it as a feature in RDS Postgres for a couple years now https://aws.amazon.com/about-aws/whats-new/2023/10/pgactive-...

But only last month did they officially release it as open source to the community https://aws-news.com/article/2025-06-09-announcing-open-sour...

Thanks, that seems to match.

A lot of theorising above, but this seems like the earliest reference

probably [Aurora DSQL](https://aws.amazon.com/rds/aurora/dsql/) which released a couple of weeks ago

I thought dsql was distributed transactions. :thinking:

DSQL does use distributed transactions to offer active-active alongside strong consistency and isolation guarantees. Here's how: https://brooker.co.za/blog/2024/12/05/inside-dsql-writes.htm...

It's definitely DSQL with the multi-region active active feature[1].

[1]https://aws.amazon.com/rds/aurora/dsql/features/#topic-1

DSQL only uses Postgres for the query processor layer, so it doesn't require a replication library within postgres itself. Definitely NOT from DSQL.

> We’re not using any of the storage or transaction processing parts of PostgreSQL, but are using the SQL engine, an adapted version of the planner and optimizer, and the client protocol implementation. [1]

Rather, DSQL seems to do its region replication using the distributed journal abstraction [2].

[1] https://brooker.co.za/blog/2024/12/04/inside-dsql.html [2] https://brooker.co.za/blog/2024/12/06/inside-dsql-cap.html

DSQL uses an internal service, Journal[1], for its mutli-region writes.

[1]: https://www.allthingsdistributed.com/2025/05/just-make-it-sc...

Yeah, and this doesn't seems to be that useful. At least I don't understand why one should do this on a strong ACID relational database.

In my experience multi-writer is because of latency or HADR stuff - have all your data in all regions at the same time, but the method (via the tlog) seems like it sort of defeats what those sorts of systems might be able to historically do (write multiple places from the app at the same time so as to have the lowest possible chance of data loss.)

Yes, I call it spoke-and-hub. The sharded spokes accept the writes and replicate back to the hub, where all shards coexist.

Useful for metric ingestion. Not useful for bank ledgers or whatever.

Yeah, some instances also have all the shards maintain all the state and just accept writes in your partition of values, merge replication in sql server works like this.

> Aurora uses it's own SAN replication layer

I don't think that is used for cross region replication

From the readme of the repo: "Use cases for this include running a Multi-Region high availability database cluster"

i think AWS is using pgactive to sell "Aurora Postgres Global" https://docs.aws.amazon.com/AmazonRDS/latest/AuroraUserGuide...

Reply on news.ycombinator.com