7 messages in com.mysql.lists.mysqlRe: How to PHP, MySQL and have secure...| From | Sent On | Attachments |
|---|---|---|
| Staci and John Morgan | 07 Jul 2000 17:55 | |
| Wesley Darlington | 08 Jul 2000 02:13 | |
| Staci and John Morgan | 08 Jul 2000 02:32 | |
| Wesley Darlington | 08 Jul 2000 03:04 | |
| Ed Carp | 08 Jul 2000 08:19 | |
| Jeremy Cole | 08 Jul 2000 10:31 | |
| Wesley Darlington | 09 Jul 2000 07:54 |
| Subject: | Re: How to PHP, MySQL and have secure login...![]() |
|---|---|
| From: | Ed Carp (er...@pobox.com) |
| Date: | 07/08/2000 08:19:21 AM |
| List: | com.mysql.lists.mysql |
Wesley Darlington (wes...@blackstar.co.uk) writes:
While your approach might be appropriate for a site with very little database-driven functionality and/or few distinct users, it wouldn't be ideal for a site with non-trivial MySQL access. Large speedups (under apache anyway) can come from caching database handles across http requests. This becomes very difficult/unwieldy when there are large numbers of website users and hence distinct MySQL users. There are OS limits on the number of simultaneous connections MySQL can support.
I have seen several people make the claim that connection caching can greatly speed up database accesses, but no one seems to be able to substantiate this claim. Is this anecdotal, or do you have any numbers to substantiate it? I seem to remember a statement Monty made a while back that said that when they implemented connection pooling for PHP, they didn't see much (if any) increase, because MySQL is very efficient at making database connections.
I would appreciate any hard data anyone can give (other than anecdotal evidence and heresay). In our testing with Escapade here in Dallas and in Florida, neither of our labs can see any significant difference, and it would be nice to be able to see if it would be worth expending the large amount of effort it would take to turn Escapade stateful.
-- Ed Carp, N7EKG er...@pobox.com 940/367-2744 cell phone http://www.pobox.com/~erc




