5 ms·

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

shagbag 2y ago
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 https://pastebin.com/aybPTd3b Please email james@mach3db.com for test credentials. Thanks!
- shagbag 2y ago
  I added the benchmarking tool to Github here: https://github.com/mach3db/postgres-benchmarker https://github.com/mach3db/postgres-benchmarker
- anonzzzies 2y ago
  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?
  - shagbag 2y ago
    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.
- anonzzzies 2y ago
  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?
  - shagbag 2y ago
    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.

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 https://pastebin.com/aybPTd3b Please email james@mach3db.com for test credentials. Thanks!

shagbag 2y ago
I added the benchmarking tool to Github here: https://github.com/mach3db/postgres-benchmarker https://github.com/mach3db/postgres-benchmarker
anonzzzies 2y ago
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?
- shagbag 2y ago
  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.
anonzzzies 2y ago
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?
- shagbag 2y ago
  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.