Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Not just that, but database performance drops quickly with writes.

1M transactions per second with 100 % reads might as well mean 10K transactions per second with 95 % readsm 5 % writes ("the usual OLTP workload" per common wisdom) and 1K transactions per second on 50 % reads 50 % writes. Or even less. Or much more. It's difficult to even quess.



What stood out to me was that the measurement was taken against 20 tables.

That makes me wonder if a smaller number of tables would have hit contention even in a read only workload.

Well it's not the only database that has this problem. I wonder if partitioning is enough or if you really need separate tables.


"The data set is 1 million rows in 20 tables. Fewer tables can be used, but below 4 tables the performance drops somewhat due to a hot spot in the table definition cache."




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: