On Sat, 24 Jan 2004, Marlon Moyer wrote: > Like Hamid Said, if the ColdFusion server has the query already in memory. > It doesn't need to send traffic to another server to get the information > again. Most systems I've worked on have the db and the web server on > different areas of a firewall, so you're going through a lot of excess steps > if you're hitting the db every time for something that won't change.
You missed the point. If you have to connect to a server anyway, whjat's the difference between going to a cached queryset on a CF server or going to a cached queryset on a db server? It's a wire transaction in any case. > Plus, you have a granular control on what is being cached. The sql server > will only cache what it has room for. So if enough queries are run against > it, the original query won't be cached anymore regardless. And this won;t happen on the CF server when you run out of memory to cache a transaction? Yeah, right. Does CF have to query the db server to stay in sync? Of course it does. Now there's multiple transactions to coordinate & synchronize. > But I think the original question was about a tree that took a long time to > create, and application variables would be a plus in this situation..... That remains to be seen. Getting the data in an optimal way can be a non-trivial operation. -- "Half the people know what they're talking about, and the other half are writing code." -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php