Hi!
I wonder if it would be possible to add some kind of caching to the driver.
Something like memory cache in pgpool:
https://lists.postgresql.org/mj/mj_confirm/domain=postgresql.org?t=06E9-8DF3-1B50
In minimal it would one more input field in driver configuration with number.
If number is 0 – driver behaves as now.
If number is e.g. 60 – then the select queries sent in the last 60s together with their results are kept in memory
And if within these 60s they are sent from the client again – the answer is return from the memory.
The rationale for this change is certain class of client applications that have some kind of ORM and issue very many
Identical queries in short time. If such are run on WAN/VPN with many network hops.
I have run some tests and currently PGSQL ODBC issues 2x more queries than MS SQL or MySQL and seems 2x slower
Because of that so these dbs are chose over PGSQL.
With such option the PGSQL should be 3-4x faster than MS SQL/MySQL.
I am aware we are losing a bit of consistency here but pgpool does the same.
Best Regards,
Jacek
