home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!wupost!usc!elroy.jpl.nasa.gov!decwrl!pa.dec.com!usenet.pa.dec.com!shah
- From: shah@tay1.dec.com (Amitabh Shah)
- Newsgroups: comp.databases
- Subject: Re: 500'000 records - who does best?
- Date: 1 Jan 93 20:45:58
- Organization: Digital Equipment Corporation, TAY1, Littleton MA
- Lines: 44
- Message-ID: <SHAH.93Jan1204558@santur.tay1.dec.com>
- References: <18774@mindlink.bc.ca> <1993Jan1.021012.24215@news.arc.nasa.gov>
- <1993Jan1.182624.2993@uunet!tellab5!odgate>
- NNTP-Posting-Host: santur.tay1.dec.com
- In-reply-to: mike@uunet!tellab5!odgate's message of 1 Jan 93 18:26:24 GMT
-
-
- On 1 Jan 93 18:26:24 GMT, mike@uunet!tellab5!odgate (Mike J. Kelly) said:
- >> lamaster@pioneer.arc.nasa.gov (Hugh LaMaster) writes:
- >In article <18774@mindlink.bc.ca>, Mischa_Sandberg@mindlink.bc.ca (Mischa Sandberg) writes:
- >|> We run Sybase on Suns and RS6000's. We chose Sybase because, within our
- >|> domain, they beat Oracle down quite handily (they also didn't try to
- >|> fudge the benchmark).
-
- >>Would anyone care to comment on the ad campaign that Oracle is
- >>running right now, showing Oracle 7 to be 3-4X faster than Sybase on
- >>TPC-A? Unfortunately, in their long list of systems compared, there
- >>is very little overlap in the system configurations run between the
- >>different vendors. The same ad has appeared in a number of trade rags.
- >>The "headline" is: Sybase Best: 183 TPS Oracle 7 Best: 645 TPS.
-
- > I would be very skeptical of any Oracle benchmark claims because Oracle's
- > license prevents the publication of any benchmark results without Oracle's
- > permission. Therefore, the only Oracle benchmark results you will see are
- > the "good" ones.
-
- While I can not comment on Oracle's policy of benchmarking in general, I can
- say something about the TPC-A benchmark results. All TPC results have to be
- accompanied by a Full Disclosure Report that gives a substantial amount of
- information about how the benchmark was run. If a vendor tries to cheat in
- their tests, chances are good that this will be caught, since their
- competitors would be closely scrutinizing the reports.
-
- That said, however, Oracle's recent TPC-A and TPC-B tests - those run with
- Oracle7 - are commonly known to use a TPC benchmark special that they
- implemented in Oracle7, viz., "discrete transactions". The latter are so
- restricted in their functionality, and in their possible interaction with
- the normal transactions, that they are utterly useless in any reallife
- application. Their sole purpose seems to be to enhance TPC-A and -B results.
- Thus, any of Oracle's claims for TPC-A or -B with Oracle7 should be taken with
- a large grain of salt.
-
- -amitabh.
- --
- ______________________________________________________________________________
- Amitabh Shah shah@tay1.dec.com----(INTERNET)
- Digital Equipment Corporation santur::shah----------(EASYNET)
- Software Performance Group +1 508 952 4430---------(PHONE)
- 151 Taylor St. Littleton MA 01460 +1 508 952 4197-----------(FAX)
-
-