In this article by Peter De Tender, the author of Mastering Hyper-V, we will talk about the building blocks for creating your virtualization platform through Hyper-V.

We need to clearly define a detailed list of required server hardware, storage hardware, physical and virtual machine operating systems, and anything else we need to be able to build our future virtualization platform.

These components are known as the Hyper-V building blocks, and we describe each one of them in the following sections.

(For more resources related to this topic, see here.)

Physical server hardware

One of the first important components when building a virtualization platform is the physical server hardware.

One of the key elements to check is the Microsoft certified hardware and software supportability and compatibility list. This list gives a detailed overview of all tested and certified server brands, server types, and their corresponding configuration components. While it is not a requirement to use this kind of machine, we can only recommend it, based on our own experience. Imagine you have a performance issue with one of your applications running inside a VM, being hosted on non-supported hardware, using non-supported physical NICs, and you’re not getting decent support from your IT partner or Microsoft on that specific platform, as the hardware is not supported.

The landing page for this compatibility list is http://www.windowsservercatalog.com.

After checking the compatibility of the server hardware and software, you need to find out which system resources are available for Hyper-V. The following table shows the maximum scaling possibilities for different components of the Hyper-V platform (the original source is Microsoft TechNet Library article at http://technet.microsoft.com/en-us/library/jj680093.aspx.)

Physical storage hardware

Next to the physical server component, another vital part of the virtualization environment is the storage hardware. In the Hyper-V platform, multiple kinds of storage are supported, that is DAS, NAS, and/or SAN:

• Direct Attached Storage (DAS): This is directly connected to the server (think of disk which is located inside the server chassis).

• Network Attached Storage (NAS): This is the storage provided via the network and presented to the Hyper-V server or virtual machines as file shares. This disk type is file-based access. Server 2012 and 2012 R2 make use of SMB 3.0 as file-sharing protocol, which allows us to use plain file shares as virtual machine storage location

• Storage Area Network (SAN): This is also network-based storage, but relies on block-based access. The volumes are presented as local disks to the host. Popular protocols within SAN environments are iSCSI and Fibre Channel.

The key point of consideration when sizing your disk infrastructure is providing enough storage, at the best performance available, and preferably high availability as well. Depending on the virtual machine’s required resources, the disk subsystem can be based on high-performant / expensive SSD disks (solid-state drives), performant / medium-priced SAS disks (serial attached SCSI), or slower but cheaper SATA (serial ATA) disks. Or it could even be a combination of all these types.

Although a bit outside of Hyper-V as such, one technology that is configured and used a lot in combination with Hyper-V Server 2012 R2, is Storage Spaces. Storage Spaces is new as of Server 2012, and can be considered as a storage virtualization subsystem. Storage Spaces are disk volumes built on top of physical storage pools, which is in fact just a bunch of physical disks (JBOD).

A very important point to note is that the aforementioned network-based SAN and NAS storage solutions cannot be a part of Storage Spaces, as it is only configurable for DAS storage.

The following schema diagram provides a good overview of the Storage Spaces topology, possibilities, and features:

Physical network devices

It’s easy to understand that your virtual platform is dependent on your physical network devices such as physical (core) switches and physical NICs in the Hyper-V hosts. When configuring Hyper-V, there are a few configurations to keep into consideration.

NIC Teaming

NIC Teaming is the configuration of multiple physical network interface cards into a single team, mainly used for high availability or higher bandwidth purposes. NIC Teaming as such is no technology of Hyper-V, but Hyper-V can make good use of this operating system feature. When configuring a NIC team, the physical network cards are bundled and presented to the host OS as one or more virtual network adapter(s). Within Hyper-V, two basic sets of algorithms exist where you can choose from during the configuration of Hyper-V networking:

• Switch-independent mode: In this configuration, the teaming is configured regardless of the switches to which the host is connected. The main advantage in this configuration is the fact the teaming can be configured to use multiple switches (for example, two NICs in the host are connected to switch 1 and 2 NICs are configured to use switch 2).

• Switch-dependent mode: In this configuration, the underlying switch is part of the teaming configuration; this automatically requires all NICs in the team to be connected to the same switch.

NIC Teaming is managed through the Server Manager / NIC Teaming interface or by using PowerShell cmdlets. Depending on your server hardware and brand, the vendor might provide you with specific configuration software to achieve the same. For example, the HP Proliant series of servers allows for HP Team configuration, which is managed by using a specific HP Team tool.

Network virtualization

Within Hyper-V 2012 R2, network virtualization not only refers to the virtual networking connections that are used by the virtual machines but also refers to the technology that allows for true network isolation to the different networks in which virtual machines operate. This feature set is very important for hosting providers, who run different virtual machines for their customers in an isolated network. You have to make sure that there is no connection possible between the virtual machines from customer A and the virtual machines from customer B. That’s exactly the main purpose of network virtualization.

Another possible way of configuring network segmentation is by using VLANs. However, this also requires VLAN configuration to be done on the physical switches, where the described network virtualization completely runs inside the virtual network switch of Hyper-V.

Server editions and licensing

The last component that comprises the Hyper-V building blocks is the server editions and licensing of the physical and virtual machines operating system.

Summary

In this article, we looked at the various building blocks for building a virtualization platform using Hyper-V.