Hacker News Reader

Comment by anonzzzies | Hacker News Reader

What to back up the claim of the fastest? For what workloads? Benchmarks vs Postgres?

I haven't conducted detailed benchmarks yet but it should be much faster than a comparable Postgres database as a service hosted on a standard cloud or colo server.

You can benchmark it with this tool: https://pastebin.com/aybPTd3b

Please email james@mach3db.com for test credentials.

Thanks!

I added the benchmarking tool to Github here: https://github.com/mach3db/postgres-benchmarker

Thanks, you are saying you are hosting Postgres on a dedi server right? That’s what you offer? Any master-master or master-slave setup or just Postgres?

You might want to change some wording on the site if that’s the case; now it looks like something more than it is I guess? I mean; taking a larger dedi server than you run would make that ‘the fastest’ no?

I offer mach3db on either a shared colo server or a dedicated colo server.

If you get multiple dedicated servers you could theoretically configure a master-master or master-slave arrangement.

The reason it's so fast is not just because it's on a dedicated server. There are actual differences between postgres and mach3db.

I invite anybody who wants to test mach3db to reach out for a free trial account. I think the speed will surprise people.

When you have, say, 100 monthly users, that are all hammering your server, how do you plan to keep it ‘the fastest’? I would think even one user can blow it up, so 2 users might require you to get another server?

If someone uses the shared server too heavily, they will need to move to a dedicated server. For any mission critical use case, the dedicated server is superior.

The shared server has 72 CPU cores so it can handle a reasonable amount of parallel activity.

Reply on news.ycombinator.com