Scaling SQLite to 4M QPS on a Single Server

Posted by exec on Friday November 12, 2021 @02:56PM
from the blog.expensify.com dept.

https://blog.expensify.com/2018/01/08/scaling-sqlite-to-4m-qps-on-a-single-server/

( 4 comments | 38 points | on HN )

Re:Scaling SQLite to 4M QPS on a Single Server by addicted on Friday November 12, 2021 @09:56PM (#29206979)

The epilogue mentions meltdown and spectre. I’m no expert on these, but wouldn’t the use case of a self hosted server where you install trusted applications only not be affected by meltdown and spectre?

My understanding was the w biggest worries with those bugs were precisely in the cloud, since they run untrusted code and run multiple users on the same hardware.

[ Parent | on HN ]

Re:Scaling SQLite to 4M QPS on a Single Server by commandersaki on Saturday November 13, 2021 @03:27AM (#29208338)

> Make RANDOM() deterministic (for some reason it’s non-deterministic by default)

What does making RANDOM() deterministic mean? Seems like it defeats the purpose of generating randomness.

[ Parent | on HN ]

Re:Scaling SQLite to 4M QPS on a Single Server by div72 on Saturday November 13, 2021 @02:21PM (#29212454)

I assume that is for the benchmark to have accurate results and not for normal operation.

[ Parent | on HN ]

Re:Scaling SQLite to 4M QPS on a Single Server by cmacleod4 on Friday November 12, 2021 @05:36PM (#29205338)

This is from 2018, but still interesting.

[ Parent | on HN ]