Re: What to do? Many updates and selects

Thiko! wrote:
> Hi
>
> Scenario 1:
>
> I am going to have a system where approx. 1000's of one line updates
> per second or one line selects per second can be expected on the oracle
> database software.
>
> The best hardware for this I guess would be critical, e.g super fast
> disks, lots of cpu etc? Oracle will be on UNIX or Linux. Most likely
> for us Linux.
>
> Scenario 2:
>
> There will be 'users' doing these updates and selects on the database
> from a web site.
> There will also be internal apps processing this data for other
> internal reasons.
>
> Should separate database used? Perhaps a replicated database could be
> used for the internal apps and querying. The internal use is simply
> SELECTs for reads. Or, the web users/subscribers will use the same
> database as the internal apps. Pros/Cons? Im thinking the internal
> apps could run at night for their select data to be collected.
>
> Many thanks for your help!
>
> Thiko!

Stick with Scenario 1 but I don't see why you think this is going to be CPU intensive. In fact I really can't buy into any of your assumptions without something to back them up. Have you put this into a test environment on a single CPU machine and tried it? What were the metrics? What did the StatsPack show?

Just going out and spending lots of money on hardware accomplishes only one thing: Spending lots of money on hardware. It might well be that the partitioning option would solve the problem but how would you know?

-- 
Daniel A. Morgan
http://www.psoug.org
damorgan_at_x.washington.edu
(replace x with u to respond)