Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: lies damn lies and benchmarks

Re: lies damn lies and benchmarks

From: crumedgeon <zimsbait_at_hotmail.com>
Date: Fri, 10 May 2002 01:12:35 GMT
Message-ID: <78FC8.77515$M7.7291775@bin7.nnrp.aus1.giganews.com>

> As a side note: SQL Server 2000 is SQL-92 ANSI compliant.
>
> The key point to remember is that SQL Server isn't a bad product. You
> may not like it because it's from Redmond and that's entirely a
> different reason.

9i is ANSI 92 compliant, also.

Regarding TPC, I wouldn't put too much faith into the results. Who knows what other variables come into play with a benchmark. Unless all databases, machines and code are exactly the same (Oracle's TPC's monitor is BEA, Compaq, IBM.. MS is always MS COM+ or MS C++) you are *still* comparing apples to oranges.

From TPC:
To facilitate an even playing field between Test Sponsors and lower the overall cost of benchmarking, the application code, database schema and definition, and tuning and optimization schemes documented in the FDR can be copied and implemented by any Test Sponsor for the sole purpose of TPC benchmarking. This precludes the reuse of TPC implementations in publicly disclosed non-TPC benchmarks.

I'm sure everyone lays out their database their own way, and coders develop code their own way. All TPC says is "this is a transaction, you design how it reacts in the system." To each their own. Received on Thu May 09 2002 - 20:12:35 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US