Re: Oracle threatens to sue Standish over an article

In article <TOM.93Jul8110608_at_amber.ssd.csd.harris.com>, tom_at_ssd.csd.harris.com (Tom Horsley) writes...
>>>>>> Regarding Re: Oracle threatens to sue Standish over an article; berenson_at_nova.enet.dec.com (Hal Berenson) adds:
>
>Thanks for a pretty reasonable exposition of this debate, but I have one
>question:
>
> Hal> 2) The real point, the one Standish made but more or less seems
> Hal> to be getting lost in the noise between them and Oracle ...
>
> Hal> (a) no longer demonstrates user-useful differences between
> Hal> products and/or (b) no longer can be extrapolated to predict
> Hal> the performance of customer applications. TPC-A and TPC-B have
> Hal> more or less reached this point. This doesn't make them bad
> Hal> benchmarks...
>
>Why doesn't this make them "bad" benchmarks? This almost seems like the
>definition of a bad benchmark. (Perhaps they were not bad at one time,
>but that isn't a reason they can't be bad now).
>--

All value of the TPC A benchmark has not yet gone away. It is still useful, for example, to compare Oracle7 performance on platform A to Oracle7 performance on platform B, as a way to measure relative hardware performance or price/performance. It has some residual value for comparing different database systems on the same hardware, but this value is fading. So I consider it obsolete. If it is still being used a year from now I'd just call it bad. (Note that I have always considered TPC-B a bad benchmark.)

.............................................................................

Hal Berenson

Home: 71640.3535_at_compuserve.com
Work: berenson_at_nova.enet.dec.com

• Disclaimer: Opinions expressed here are my own and should not be construed as representing the views of my employer or its employees, officers, directors, or stockholders. --