Some folks are getting annoyed when the results generated by the newly released Exchange 2013 server role requirements calculator don't match the results from the Exchange 2010 calculator. I don't know why this might be so, but it is. More positive news comes in the release of Exchange 2010 SP3 RU1 and the 26 bugs that it fixes plus one useful backport to make the lives of those looking after hybrid deployments a little easier.
I guess that it was inevitable that people would rush to use the Exchange 2013 server role requirements calculator and compare its output to the results generated by the Exchange 2010 mailbox server calculator. And of course, the results do not match. Not by a long way, because uses a different architecture than its predecessor and because server hardware has different capabilities than it had when Exchange 2010 was designed.
Microsoft did their best to make it clear that the two calculators are based on different data and make different assumptions, noting in an EHLO blog post that “the megacycle guidance in Exchange 2013 leverages a new server baseline, therefore, you cannot directly compare the output from the Exchange 2010 calculator with the Exchange 2013 calculator.” That note seems to have been completely missed by some.
Major releases of Exchange have inevitably used different hardware configurations. It has been so since Exchange 4.0 appeared in 1996. You wouldn’t take the results generated for a Digital Alpha server running Exchange 4.0 on Windows NT 3.51 and compare them against an Intel Pentium server being prepared to run Windows 2000 and Exchange 2000, even with the same number of mailboxes. It therefore baffles me why people expect to take the results generated for Exchange 2010 and be able to make much sense when compared to results for Exchange 2013 when the new product uses a completely different implementation of the Information Store, a dramatically different (and more memory-hungry) search facility, offers more features and functionality, and runs on different server hardware.
I’m not saying that you should take the results generated by the calculator and use them to create an order for your preferred hardware vendor. Doing so without engaging your brain from neutral into at least first gear would be stupid as you absolutely have to query any result from a generic sizing tool and put it into the context of your company’s business, operational, and technical requirements. It is only then that you can arrive at a solid decision as to an appropriate hardware configuration.
In other news, Microsoft released Exchange 2010 SP3 RU1 on May 30, 2013. KB2803727 provides an outline description of the 26 fixes included in this roll-up update, the most important of which include the fix for the soft-delete problem that afflicted users who work with Outlook in online mode. One little gem that is slipped into RU1 is a back-port from Exchange 2013 that updates the Set-HybridConfiguration cmdlet to support multiple domains that are supported for Autodiscover, a very useful enhancement when you are faced with the need to support users with email addresses for more than a principal domain in a hybrid on-premises/ deployment. Like any other update, please test this software before deploying into production – Just in case.
In closing, I would like to announce that my new pet hate is the misuse of the term “on-premises”. Those who insist that the correct usage is “on premise” are incorrect. Sorry. Please remember that you work inside a premises, unless of course you work on a premise (“A proposition upon which an argument is based or from which a conclusion is drawn.”) There is a difference you know, just like a difference in meaning exists between the words “principle” and “principal”, as all you principle senior consultants or whatever-you-are should well know.
Follow Tony @12Knocksinna