Exchange 2010 Mailbox Server Role Requirements Calculator v6.3 Released

What do I think ? I think Ross is VERY clever, and I'm very gateful for such a useful tool! Thank you Ross and everyone involved.

"This workbook contains content that is not supported by Excel 2008 for Mac.
* Structured references
* Excel tables
* Table styles"

Is it worth someone engaging the MacBU to see how to best make it compatible across the board?

Hi Ross, I think this calculator is great thanks for putting so much time and effort into it.  A couple of quick questions, 1st is about the megacycles example calculation, is the G5 used as the base because it is an industry standard or should we be using the base value found in the SPEC site for the server we are looking at?  In the calculator it uses the DL 380 G5 x5470 3.33 GHz processor for a base of 18.75 but the processor in question is a  x5550 at 2.65 Ghz.  Should the DL380 always be used as the base value?

The second question I have is, when we work the megacycles numbers is the fact that a proc is hyper threaded already accounted for or can we safely boost our megacycle numbers by a factor of "X" i.e 1.5?  Thanks again.

John

Hey John, all of our megacycle / message profile calculations were done against the DL380 G5 platform I mentioned in the note, thus why you need to do the steps I outlined in the comment.

I'll need to get back to you on the second question.

Ross

I thought the Exchange team still suggested Hyper Threading be disabled as a best practice.

John, we cannot accurately predict the impact of hyper-threading across all hardware platforms. All of our guidance is based on running with hyperthreading disabled. We recommend that you do your initial sizing with hyper-threading disabled. If you choose to run with hyperthreading enabled then you should be doing your own validation testing to determine the impact.

Ross

Great tool! One question, though: The megacycles adjustment is defined to made using the formula Megacycles/core = ((New platform per core value) * (Hertz per core of New platform)) / (Baseline per core value). Shouldn't this be Megacycles/core = ((New platform per core value) * (Hertz per core of baseline platform)) / (Baseline per core value)? At least for me, it would make sense that the adjustment is made based on performance in relation to the baseline system.