No, not really.

For RBS, we all have rules of thumb. Its an empirical thing.

See
</code> http://asktom.oracle.com/pls/asktom/f?p=100:11:::::P11_QUESTION_ID:233214228495

for the details.  I myself like to start with about 15 or so (depending on the number of concurrent transactions) rollback segments -- each with initial=next=1m and minextents=25, maxextents=100.  So I have 15 to 25 25meg rollback segments that can grow to 100meg each.  I'll shrink them before backups back down to 25meg.

For tables, then only real way to do this is to load some sample data. See

http://asktom.oracle.com/pls/asktom/f?p=100:11:::::P11_QUESTION_ID:302216525000 <code>
it also points to some doc that has formula's, but since everything is a varying string pretty much -- they are usually pretty far off from reality.


Log files - you size and resize (not really resize but create new and drop old) as part of tuning. the size of your redo affects performance more then anything else (eg: having too small redo will not prevent you from doing anything like having too small rollback would).

I like to start with (depending on the system) 3 25meg log files. This is for "moderate" systems. The goal (some people will say) is to have sufficient online redo log to get you through your busy period without wrapping around for about 15-20 minutes. Their size depends on how frequently you want to checkpoint and how large you want them. I've used 25meg, 100meg and 255meg. You can go larger but I've not done so.