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: A performance question: Unix v/s NT

Re: A performance question: Unix v/s NT

From: Jerry Gitomer <jgitomer_at_p3.net>
Date: 1997/08/24
Message-ID: <34011273.2C12@p3.net>#1/1

otierney_at_freenet.edmonton.ab.ca wrote:
>
> performance (by which I mean speed on a large system, ie tpcc on a
> system of large size) and reliabilty are seperate if (vaguely) related,
> in my humble (biggoted/biassed towards unix) opinion:
>
> Performance;
> Microsoft Windows NT does not scale, it is related to (though a complete
> rethink/rewrite of) a single user, single task system (msdos5/windows3.0).
>NT is based on the Mach Kernel (which will be the basis of the OSF
operating system that will probably replace UNIX) and the key NT developers were lured away from DEC's operating system group. The user I/F is based on Windows 3.1 (prior to NT 4.0) and Windows 95 (beginning with NT 4.0) but internally it is a real operating system.

As far as scaling goes I am in the process of configuring an NT system using the Oracle Parallel Server option which will have a total of 16 processors (4-4 processor nodes) accessing a 4-port RAID box with one controller per spindle. I anticipate mainframe level performance.

> Reliabilty;
> If you can control your system in order to run your application and system
> to run as few new (ie buggy and unreliable) bits of the system as
> possible.
> My two biggest concerns about my proposed configuration are reliability
of NT and the quality of the support that I will be able to get from Microsoft when I encounter problems.  

<snip>

Jerry Received on Sun Aug 24 1997 - 00:00:00 CDT

Original text of this message

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