I agree a lot with you. We have Oracle on 3k since two years. At first,
since I didn't kow well Oracle we had a few problem, but now it performs
quite well. Moreover, I have installed Oracle on NT and I can tell 3k is a
better platform to tune Oracle than NT (my own opinion, of course). We also
had a few problems with SQL*Net, but almost not with the DBMS itself.
I still think that Oracle could be a good friend for the HPe3000. And at
some point, I wonder if Oracle is a "must be" friend for any platform...
Michel Gauthier
Donohue Inc.
Joe Smith <[log in to unmask]> a écrit dans le message :
F4B1826B1A21D211AEC5006008207AF404377A8F@dogbert.csillc.com...
> The "need" for Oracle components or its DBMS on the HP3000 should be
viewed by
> HP as an opportunity to sell hardware, and to further expose the e3000 to
an
> enormous market. Oracle performs so well in the HP3000 environment that
HP has
> a solid winner when choosing platforms, if they would dare to promote
this.
>
> We have been running Oracle on our HP3000 platform for 4 years now.
Other than
> a network services / Oracle transport layer induced system abort problem,
Oracle
> has run flawlessly. We currently have 32 Oracle applications and about
140
> concurrent Oracle users on our systems - an HP987. Average cpu is about
60%
> during peak usage, this includes TurboImage applications. The response
time is
> excellent - given the size of the box against what it has to run.
>
> Let's not start a thread here about the virtues of TurboImage vs Oracle -
I
> fully agree that Oracle costs many times more, is more difficult to learn,
takes
> more people to manage, takes more horsepower to run - been there, doing
that.
> The problem we are facing is the direction our company has taken, along
with
> many others, to purchase external packages and run them on Oracle.
Without
> exception, everything our apps people bring in to look at run on Oracle.
Not
> one has TurboImage as a possibility. Even Big Iron is losing out in a big
> way....
>
> The days of Oracle on the HP3000 are coming to a close for us. Last year,
we
> started to see a number of Oracle applications require 8.0 or higher to
run.
> The highest version supported on the HP3000 is 7.3.4 with no plans to go
> forward. Thus, a decision was made to bring in large 4 way NT servers to
run
> the 8.0 apps. We got up to eleven 4 way servers running a total of 15
apps,
> some seeing response times of 3 - 4 minutes...our DBAs hate the admin
aspects of
> this platform.
>
> Our next solution, which we are now beta testing, is an HP9000 N class
with 2
> 440 processors and 5 gig memory. Initial results show a slight
improvement in
> performance. Although HP still sold hardware to run Oracle, they could
have
> easily have lost out to Sun. Sun is a very acceptable platform for Oracle
and
> should see at least half of the business in this area. The only reason we
> choose HP is the acceptance of HP products in our company (given all else
> equal). This acceptance, ironically, was fostered by the reliability of
the
> HP3000!
>
> I'm sure HP stopped putting money into Oracle on the HP3000 when they
didn't see
> a profitable return. Perhaps if they advertised success stories to the
Oracle
> world, showed the cost / performance benefit, and given more time they
could
> have hit the critical mass to make a profit. By stopping development long
ago
> at 7.3.4, the possibility of widespread acceptance of the HP3000 as a
viable
> Oracle platform may not be possible. Certainly for us, we will migrate
our
> HP3000 Oracle apps and NT Oracle apps to Unix. One less reason to keep
our
> HP3000...
>
|