If you want to be buzzword compliant, call your system a 'data warehouse.' Commercial database companies will sell you this same type of solution starting at six figure USD prices. I sat through a long sales pitch. As far as I could tell, your idea is the only database improvement involved in data warehousing.
I was able to get the mod_perl processes to share the memory used by the Storable data structure. This way the many http daemons didn't consume too much RAM. We had three data structures using a total of about 100M bytes of RAM.
You have to set the operating system process size limit large enough so that you don't run out of memory for the data structure.
One very minor gotcha we ran into was that the Storable data was OS dependent. It wasn't a network order problem; we just upgraded our web server OS and somehow it wouldn't read Storable data from the old OS any more. It shouldn't have had that problem, so there was probably a bug somewhere.
It should work perfectly the first time! - toma
In reply to Re: mySQL hits or Storable retrieves?
by toma
in thread mySQL hits or Storable retrieves?
by Hero Zzyzzx
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |