Re: multi-user database design

Generally, invoking a connection to a database is pretty slow. So you want some way to "multiplex" that connection.

So, yes, one uses something like Binary Evolution's Velocigen Engine for Perl, mod_perl, etc., to not only minimize the number of perl processes invoked (and to *not* use CGI to call Perl), but also to minimize the number of database connections per web server (and use many web servers). One lets multiple web servers be the multi-headed hydra interfaces to the web, and each hydra's tail provides the conduit to the database server...

Of course, sites like Amazon.com, with all that, still use about as powerful of a database server they can to handle hundreds or thousands of db transactions per second...

<racudd_at_my-deja.com> wrote in message news:862ut1$19f$1_at_nnrp1.deja.com...
> How do the likes of Amazon and Hotmail etc (anyone with loads of users)
> design their databases ? Would they set up a new database user for
> each of their customers and thus have literally thousands of users
> accessing their database at a time. Or would they have one or two
> "power" users to access their databases and retrieve all of the data
> and then pass it back to each individual user ?
>
> Also, would they use persistent db connections (with Apache and
> mod_perl or something) to increase speed or would they try and save
> some memory and just make a new connect to the database each time the
> users request information? I can imagine that persistent connections
> would limit the number of concurrent users possible due to memory usage.
> Thanks
> Richard
>
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
Received on Wed Jan 19 2000 - 00:21:37 CST