Since fragmentation need never happen these days, that leaves "extent
lifespan". Quite how an index extent would live longer or lesser than its
corresponding table extent beats me. Maybe you could elaborate.
HJR
"simon" <onlysimon_at_hotmail.com> wrote in message
news:cs7lcuoue1of1bsg7462vlbtqhgpfi5gmb_at_4ax.com...
> On 21 Apr 2002 11:08:16 -0700, Thomas Kyte <tkyte_at_oracle.com> wrote:
>
> >
> >
> >The theory behind putting index and data on separate devices goes like
this:
> >
> >"you read the index, then you read the data. We'd like to leave the head
> >positioned in the index structure just so -- so that the next read on it
doesn't
> >require head movement". (or something like that).
> >
>
> I don't think that is the case.
>
> Separate DATA and INDEX tablespaces because of fragmentation and
> extent lifespan.
>
> You could put these tablespaces on separate disks to spread disk I/O
> in a MULTI USER environment.
>
> It would be even better to spread both tablespaces over different
> disks to spread disk I/O.
>
> Simon (OCA)
Received on Sat Apr 27 2002 - 12:04:54 CDT
