Thank you Gail for the link.

From the link:

'summary would be: don't change or implement anything without research and without knowing what it's going to affect. '

My quest at the moment is introducing Clustering for a number of tables. By now I have shown that clustering will give an significant improvement for allmost all actions. (Select/Insert/Update/Delete, complex statements, cache). And no significant degradation for any action.

But I do not have the resources to benchmark all situations. (Partitioning and thousands of users is one of them and a enormously large number of potential statements).

But the company is 'a bit' reluctant to the change. So I am trying to gain some knowledge about partitioning, do this by reading articles (yours) and by performing some simple 'tests', hence my question. With the understanding of internals I am in a stronger position.

Thanks for the link,

Ben Brugman