performance issue

hi,

in our application we have a screen that shows the state of activities for
each customer we have.
it's like:

company | act1 finished? | date of act1 finished | act2 finished? | date of
act2 finished | ....

i guess you can imagine what i mean with this.

problem is: the list includes hundreds of customers, and we excpect to be
that thousand soon. so what our code does is:
- get a list of all customers (which contains process references)
- query the status of specified activities in each process through shark API
- put the customer data and the results from the queries to shark into a
list

we have A REALLY FAST machine here, and it takes 50 seconds. before we
started to query shark for the activity states, it took 0.06 seconds in
average. we were aware of this problem, but we still do not have a solution
that fits both architectural and performance requirements.

to sove the performance problems, we thought about querying the shark
database directly without the API. as we do not change state, that shouldn't
be a problem. but of course we can get into troubles when the shark db
layout changes.

we would like to stick to the API, but do not see a way to overcome the
performace issues. maybe indicies? but if so, which columns shuld be
indexed?

so my questions are in detail:
- is there a chance to get a database documentation for the DODS
implementation (like charts including FK-PK relations, etc.)?
- anyone an idea how to make the API calls faster?

kr,
guenther


--
Günther Wieser

creative-it
Guglgasse 6/1/11/1
A-1110 Wien
Austria
http://www.creative-it.com