| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Usenet -> c.d.o.server -> Re: Windows vs Linux Performance
DA Morgan wrote:
> Charles Hooper wrote:
>> hpuxrac wrote:
>>>> >>>> Matt >>> Looks like it's based on itanium ... I started laughing and didn't read >>> any further. >> >> The results were not what I expected, and may have been very different >> given a slightly different hardware configuration. It does, however, >> show that RAC works with Windows platforms, at least in test >> environments. >> >> Charles Hooper >> PC Support Specialist >> K&M Machine-Fabricating, Inc.
aim of the test is not RAC, from the intro it is about providing at least some evidence.
> 2. You rely on an undocumented configuration
Not true, you could argue that it is insufficiently documented since it
only references the hardware config in the document itself and
references the OS and Oracle installs in the appendix. But the build is
documented.
> 3. You rely on an undocumented test suite
I take it you didn't bother visiting dominic's site which has the source
downloadable?
> 4. You don't mind that Intel codes pathing into their chips that
> favors Windows over Linux.
This is the second time to my knowledge that you have asserted this as a
bad thing and relevant to Oracle database performance. Surely it's
either irrelevant FUD or another argument in favour of Windows as an os
on intel hardware.
> 5. You ignore the dangers from rootkits, viruses, trojans, etc.
firewalls and good administration. The fact that my open source website got hacked recently despite being on a firewalled, protected linux install without root privs is at least evidence that the problem isn't os specific, oh yes I could have avoided it maybe if I'd kept up with the constant stream of security fixes from the open source vendors. Hang sounds a bit like ms and oracle that.
>
> For the rest of us there is a real operating system.
and of course unsubstantiated and unreproducible assertions such as 'on linux it is sub second on windows it is 25 second' where the test, the hardware and so on are not even cursorily mentioned. Or perhaps you intended people to read VMS.
I've probably gone overboard with my response, but every time I have seen comparable installs tested on the two OS's the results have been, well comparable. For the most part however people choose to impune the motives of the testers, suggest incompetence on the part of the testers or else just offer practical irrelevances (supposed code path bias towards windows from Intel for example). If people offered documented tests showing different results that would be one thing, but anecdotes and aggression in the place of evidence and argument are not generally a great sign of good thinking.
And yes, my position remains that technically the two platforms are roughly equivalent so make decisions based on other, primarily, human factors.
-- Niall Litchfield Oracle DBA http://www.orawin.info/servicesReceived on Wed Jan 24 2007 - 14:16:43 CST
 |
 |