Hi,

> Well I was not trying to *blame* the hardware; I was actually thinking
> that it's the real reason; I can tell you what kinds of queries I'm
> talking about and maybe you tell me whethere I'm doing anything in a
> way that it could be done better:

having looked over your post i have two questions:
- Why don't you use FIRST/SKIP syntax and instead submitting something
like a starting ID. I've used that in many applications and it usually
runs pretty fast.
- Have you tried looking at the PLAN's of the queries you do?
Understanding PLANS's and adding the approriate indexes is the key to
speeding up things considerably.
- If all fails, you could just define something like "results table"
which includes exactly the columns you expect from your query plus a
"query id" column. You would generate this "query id" every time a user
requests results and fill your results into the "results table" with the
same "query id". Then you should be able to easily navigate very fast in
this table via primary key and "query id". The user would only have to
wait for the results the first time, not after paging. I hope this
doesn't sound only glibberish to you ;)

kind regards,
Michael