Re: Implementing "thick"/"fat" databases

On Mon, Jul 25, 2011 at 1:24 AM, Sim Zacks <sim(at)compulab(dot)co(dot)il> wrote:

>
> If I understand you correctly, you are saying that to handle business logic
> processing, I may require X servers. Only a percentage of that traffic
> actually requires database processing. if I use a cluster of application
> servers against a single database, it will scale better then if I have to
> cluster my database, which brings in all sorts of messy master-master
> replication issues.
>
> Is this accurate?
>

I'd also say I have seen applications not put enough logic in the
database, and that this can cause worse bottlenecks.

Last time I looked at SQL-Ledger's payment processing interface, it
did in Perl in the web app what would have been far better solved with
a HAVING statement in the SQL (retrieve ALL invoices ever entered,
along with aggregated payments and then compare these numbers in Perl
and discard those which are already paid).

Needless to say, this approach, when run on a large database, caused
DB and web server bottlenecks.......

Doh!

Best Wishes,
Chris Travers