Specifications

Best Practices for Virtualizing and Managing Exchange 2013

Guest Storage

In addition to presenting VHD or VHDX files to Exchange 2013 virtual machines, administrators can

choose to connect the guest operating system of an Exchange virtual machine directly to existing storage

investments. Two methods provided in Windows Server 2012 Hyper-V are In-Guest iSCSI and Virtual Fibre

Channel.

In-Guest iSCSI

Deploying an Exchange 2013 virtual machine on iSCSI storage provides a more cost-effective solution for

enterprise-level virtual machine installations than an equivalent Fibre Channel solution. Instead of using

virtual disks (such as the VHD or VHDX files discussed earlier) and placing them on the iSCSI LUNS

presented to the host, the administrator can choose to bypass the host and connect the virtual machines

directly to the iSCSI array itself. The iSCSI target, which is part of the storage array, provides storage to the

Exchange 2013 virtual machine directly over the virtual machine’s network adapters. The Exchange virtual

machine uses the in-box iSCSI initiator inside the Windows Server guest operating system to consume the

storage over a vNIC that has connectivity on the iSCSI storage network. The respective Exchange Servers

can therefore store information, logs, and other critical data directly on iSCSI disk volumes.

To enact this approach, the administrator must create dedicated Hyper-V virtual switches and bind them

to appropriate physical NICs in the hosts. This ensures that the virtual machines can communicate with

the iSCSI storage on the appropriate network/VLAN. After configuration, the administrator must use the

guest operating system IQN from the iSCSI initiator to present the appropriate LUNS directly to the virtual

machine over the virtual networks. In addition, vNIC features like jumbo frames and some other offload

capabilities can help to increase performance and throughput over the network. It is important to note

that if you intend to run the virtual machine with In-Guest iSCSI on top of a Hyper-V cluster, all cluster

nodes must have the same iSCSI virtual switches created on the hosts to ensure that when the virtual

machine migrates around the cluster, connectivity to the underlying storage is not lost.

For resiliency, the administrator may want to use multiple vNICs to connect the virtual machine to the

iSCSI SAN. If this is the case, it is important to enable and configure MPIO, as discussed earlier, to ensure

optimal performance and resiliency.

Virtual Fibre Channel

In a similar way to iSCSI, Virtual Fibre Channel for Hyper-V helps to connect to FC storage from within a

virtual machine, bypassing the host’s operating system. Virtual FC for Hyper-V provides direct SAN access

from the guest operating system by using standard World Wide Node Names (WWNN) and Worldwide

Port Names (WWPN) associated with a virtual machine. Virtual FC for Hyper-V also helps to run the

Failover Clustering feature inside the guest operating system of a virtual machine connected to the

underlying, shared FC storage.

For virtualizing Exchange 2013, Virtual FC for Hyper-V allows you to use existing FC investments to drive

the highest levels of storage performance access, while also retaining support for virtual machine live

migration and MPIO.