Re: >Oracle's future

From: Surendra K. Verma <verma_at_eclipse.torolab.ibm.com>
Date: Thu, 24 Mar 1994 19:13:29 GMT
Message-ID: <Cn6o2H.19H5_at_eclipse.torolab.ibm.com>


In article <ddargo.764466030_at_ibmsun1> ddargo_at_us.oracle.com (Dave Dargo) writes:
>verma_at_eclipse.torolab.ibm.com (Surendra K. Verma) writes:
>
>>In article <1994Mar22.223454.2137_at_colesmyer.com.au> amay_at_colesmyer.com.au (Andrew May CMIS) writes:
>>>>I would just like to point out several things:
>>>> 1, DB2/6000 is faster than Oracle v7, if Oracle doesn't use discrete trans.
>>>
>>>You would hope that IBM's proprietary database on their proprietary box
>>>would run faster than any other. The figures I got from IBM are:
>>>
>>I don't know how you can classify DB2/6000 to be any more "proprietary"
>>than Oracle. Just because it was released on the RS/6000 before any
>>other platform? I think what's great about the statement below is that
>>the FIRST release of this product would have beaten Oracle if Oracle
>>had not used discrete transactions.
>
>First, DB2/6000 was not released on the RS/6000 before any other platform.
>DB2/6000 is IBM's Workstation RDBMS and is a port of the DB2/2 product.
>It was released on OS/2 first and is just the next release of the OS/2 EE
>Datamanager.

Sorry, I meant first UNIX platform. Again, DB2/2 is more enhanced than OS/2 EE Datamanager. For example, the log-manager has been rewritten. DB2/2 is also much more stable than OS/2 ES1.0 which came after OS2/EE. This is something that our customers tell us. DB2/6000 is even more enhanced in terms of stability, performance, and function than any of the OS/2 products. I'm personally responsible for contributing quite a bit to the performance of DB2/6000.

>
>Secondly, on what do you base the statement that it would have beaten Oracle
>if Oracle had not used discrete transactions? A guess? A feeling? A hope?
>
>Dave (ddargo_at_us.oracle.com)

Why don't you tell us how much you'd get if you didn't use discrete transactions?

I'm making the statement on the basis of the "widely known" fact that discrete transactions buy about 30% on TPC-A. You can do the calculations.

Surendra
verma_at_torolab2.torolab.ibm.com Received on Thu Mar 24 1994 - 20:13:29 CET

Original text of this message