A large portion of my work consists of designing Exchange 2010 environments for customers all around the world.
Recently I was working on an assignment where we wanted to implement Exchange UM to replace cisco unity for 23 offices. The whole deployment is planned to run virtualized, this directed me to the specific requirements of UM on virtualization.
This project made it once again clear to me that some good research is valuable even when you think you know your stuff.
What I already knew was that Exchange UM was only supported to be virtualized when it was running on its own, as well as that it required 4 processor cores (with no oversubscription).The requirement that I wasn't 100% sure about, was the memory.
So while doing research on this, I found some contradictory information on this.
If you read the Best Practices for virtualizing Exchange 2010 whitepaper
http://www.microsoft.com/download/en/details.aspx?id=2428
You will notice that the requirement besides 4 Processor cores would be to allocate 16 GB of RAM.
Though when you follow the guidance from the Exchange help at
http://technet.microsoft.com/en-us/library/aa996719.aspx
You will notice the requirements state the following text:
All Exchange 2010 server roles, including Unified Messaging, are supported in a virtual machine. Unified Messaging virtual machines have the following special requirements:
- Four virtual processors are required for the virtual machine. Memory should be sized using standard best practices guidance.
Looking at the standard best practices guidance it will tell you the minimum requirements is 2GB per processor core, doing the calculation for you it makes 8GB as a minimum requirement instead of 16GB.
So I sent out a mail to check what the official support standpoint was, the result is now that the best practices whitepaper has been updated to reflect the guidance from the Exchange help.
So the final requirements for running Exchange 2010 UM on a supported virtualization platform is 4 processor cores (with no oversubscription) and 8 GB of memory.
In my design this saved the customer 184 GB of memory, worth spending some time on.