Re: tx isolation

From: Brian E Dick <bdick_at_cox.net>
Date: Mon, 02 Dec 2002 14:02:40 GMT
Message-ID: <4KJG9.65122$wc2.2718284@news2.east.cox.net>

One of the biggest shortcomings of the TPC-C benchmark is its "shared nothing" architecture. It allows you to infinitely scale the benchmark without concern for resource contention. By side-stepping the locking issue, this benchmark gives the SQL Server locking model an unfair advantage over the Oracle locking model. In a real world "shared something" or "shared mostly" application, Oracle shines.

Also, multiversioning does not make for lazy developers. Multiversioning resolves a difficult problem, resource contention, and allows the developer to move on to bigger and tougher problems.

"Pablo Sanchez" <pablo_at_dev.null> wrote in message news:Xns92D7F0C77F0Fpingottpingottbah_at_209.242.64.107...
> KevJohnP <nospam_at_nowhere.com> wrote in news:3DEAC080.8020402_at_nowhere.com:
>
> > In the interests of balance, tried to find some Microsoft links that
> > refute this.
>
> Lock escalation issues are ... non-issues. Multiversioning is great
> but it also makes for lazy developers and the industry database
> benchmarks don't show that Oracle has any clear advantage.
>
> If one takes a gander at the current (12/1/02 @ 11:38pm MT) top 5
> clustered TPC-C by Performance, Version 5 results, the top three spots
> are owned by SQL Server 2000. Oracle 9i R2 takes the 4th and 5th
> spots.
>
> SQL Server 2000 tpmC: 709,220 <<<<----- Yeow!!!!!
> Oracle 9i R2 tpmC: 138,362
>
> One can look at the non-clustered numbers and there, Oracle 9i R2 looks
> more impressive. Nonetheless, SQL Server 2000 is up there as well.
>
> Bottomline: multiversioning is all marketing and it eats resources to
> boot.
>
> Does it do the dishes too?
> --
> Pablo Sanchez, High-Performance Database Engineering
> http://www.hpdbe.com
Received on Mon Dec 02 2002 - 08:02:40 CST