Download the Microsoft Exchange Calculator here: http://aka.ms/E2013Calc
Read the "Microsoft Exchange Calculator Readme" for comprehensive guidance on using the calculator, here: http://aka.ms/exchangecalc
The focus of this Sizing exercise is on the parameters that should be configured on the Input Tab of the Microsoft Exchange Calculator. Customers that plan to deploy in multiple datacenters and stretch the Database Availability Group (DAG) should ensure to complete the input for the secondary datacenter, under Site Resilience Configuration, so that the primary datacenter Compute and Storage requirements are properly sized to handle all users in the event where the secondary datacenter is down.
Figure 1) Exchange Environment Configuration
If virtualizing Exchange, ensure that “Server Role Virtualization” is set to Yes. If a DAG is planned, ensure “High Availability Deployment” is set to Yes, and that the proper number of database copy instances are selected for each site.
Figure 2) Tier-1 [2, 3, 4] User Mailbox Configuration
Ensure the user mailbox tiers are set to the appropriate initial and maximum mailbox sizes. Keep in mind that the total amount of storage required will be based off the 'Mailbox Size limit'. With Pure FlashArray the storage can be expanded to the controller maximum without disruption and since everything is thin provisioned, expanding Pure Volumes is easy. Adding Pure Volumes, additional Databases, or expanding the Windows LUN where a database is located, is instant and can occur without any downtime. Keep this in mind when selecting the 'Mailbox Size limit' so that it is an estimate for the not too distant future, limiting the amount of storage purchased today and not utilized for a long time.
Figure 3) Backup Configuration
Backup Methodology can have a sizing impact. The recommended methodology is to utilize the Pure Hardware VSS Provider to enable an application consistent snapshot. 3rd party backup application that will keep a copy of the backup off the cluster, typically in a backup repository, are also valid with and without Pure integration. That is shown below as the “Hardware VSS Backup/Restore”, which will require the smallest amount of capacity for restore LUNs. Both “Software VSS Backup/Restore” and “Exchange Native Data Protection” will require restore LUNs much larger in size. If the customer is utilizing VMware redo-log snapshots, or a combination of “Exchange Native Data Protection” with lagged database copies, large restore LUNs must be provisioned on each Exchange Server to provide enough space to pull a copy of the backup and then enable roll forward recovery.
Figure 4) Storage Options
If "Automatically Calculate Number of Exchange Database Volumes Required" is set to No, carefully size and ensure enough Exchange Data Volumes are selected so that all of the databases will fit on the server.
Figure 5) Server Configuration [C220 M5 Xeon 8170 2.1GHz SPECint2006 example]
The Microsoft Exchange Calculator is based on a particular baseline CPU. In order to property calculate the megacycles to actual CPU consumption fill out both the number of vCPUs for the Exchange Server VM in “Processor Cores / Server”, and the “SPECint2006 Rate Value” for the Server.
Select the number of processor cores in the server, or assigned to the VM. Next go to spec.org and find your server and CPU, or at least the exact CPU in another model of Server, and note the "SPECint_rate2006 Result" value. For Physical Exchange Servers, this value is what you enter. For Virtual Machines the following formula can be used to identify the value.
X / (N*Y) = per virtual processor (core) assigned to VM
X = SPECInt2006 rate value
N = Number of physical cores on the hypervisor host
Y = 1 if 1:1 vCPU:CPU ratio and 2 if 2:1 vCPU:CPU ratio
The example C220M5 with the Xeon 2.1GHz in Figure 5 has a SPECint2006 rate value of 2430. If assigning 16 cores out of the 52 on the server to the VM, the formula would be as follows:
2430 / (52 * 1) = 46 per core
16 cores * 46 = 736