[nycbug-talk] Lab environment

[G.Rosamond]

On May 19, 2004, at 10:47 PM, Isaac Levy wrote:
>> But what are we comparing then?
>
> Well, as I expressed to Sunny just 1 min ago, I believe that a raw 
> performance test is as foolhearty as comparing X86 to PPC to Sparc to 
> Power4 architectures- there is really no clear 'winner', only 
> strengths and weaknesses...
>

Yeah, but mine was first. . .<g>

And as I said, we have to keep some controls. . .one platform i386 
preferably.  We should do something like one OS (ie, a BSD) and all the 
relevant network protocols.  And then there's the question of tweaking 
OS's, which is frequently an issue in every test, particularly with the 
BSDs and Linux. . .



>>
>>>>
>>>> Thoughts?
>>>>
>>>> Who is interested in taking a lead on this?
>>>
>>> I don't know about lead, but count me in.  If it's a lead that's 
>>> necessary, I'll do it, but not until after next NYCBUG meeting can I 
>>> commit the time.
>>
>> I take that as volunteering. . . <g>
> [...snip...]
>> We need to create a test environment, so we have to have each 
>> variable controlled.  Not to sound like I just took 6th grade 
>> science. . .
>
> Ahh- seeing as you have a good idea of the network enviornment, I'd 
> say I can't go this one alone, but will need some backing on how the 
> hardware/network/softwares are setup and controlled.
>
> So I'll keep volunteered if some folks get my back on this stuff... <g>

I think the parameters need to be elaborated on the list, and then we 
can worry about the setups.

g