Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.server -> Re: separate data/inidex
I suspect this came about with Oracle 5
when disc was expensive, memory was
tiny, and every Oracle block access was a GET
with no options for a PIN, so that logical I/O
was converted to physical I/O at a much higher
rate than it was with newer versions of Oracle
and larger, modern machines.
Even then it was daft because it encouraged DBAs to put one table in tablespace X and its N indexes in tablespace Y - thus insuring that a single row inserted would require one block write on one disc, and N block writes on another.
-- Jonathan Lewis http://www.jlcomp.demon.co.uk Author of: Practical Oracle 8i: Building Efficient Databases Next Seminar - Australia - July/August http://www.jlcomp.demon.co.uk/seminar.html Host to The Co-Operative Oracle Users' FAQ http://www.jlcomp.demon.co.uk/faq/ind_faq.html Connor McDonald wrote in message <3CC2F8F8.7D93_at_yahoo.com>...Received on Sun Apr 21 2002 - 15:17:53 CDT
>
>Separating data and indexes is a very (very) special case of the
>argument for balancing IO. I would suspect that it came about
>originally when everything when running rule based optimizer so your app
>was doing full scans or nested loops and that was it...Even then the
>advice was dubious in a multiuser database anyway...
>