Implementing a Hyper-V Virtual Network

Windows Server 2012 Hyper-V enables the creation of virtual networks to connect virtual machines (VMs) that can coexist on the same physical network despite duplicate IP addresses. The illustration of Network Virtualization Generic Routing Encapsulation (NVGRE) shows two physical servers hosting VMs for two organizations (Org 1 and Org 2) with two virtual networks in a multitenant environment.

Each VM shown in the top row has a virtualized network interface card (NIC) to which the organization has assigned an IPv4 or IPv6 address (or both). On a virtual network, this organization-assigned IP address is called the Customer Address (CA). (To simplify the illustration, the dot-decimal notation of all IPv4 addresses has been shortened; for example, 10.0.0.05 is shown as .05.) In addition, all VMs that communicate on the virtual network are assigned a common Virtual Subnet ID (ID 5001 or ID 6001 in this illustration).

The physical NIC in each physical server (bottom) is also assigned an IP address. As this address is typically assigned by the service or network provider, it is referred to as the Provider Address (PA). (Again, to simplify, the IPv4 IP address has been shortened; for example, 192.168.4.11 is shown as .11.) The combination of the IDs and the CA and PA addresses allows multiple virtual networks to be created on the same physical network, while securely isolating traffic to the appropriate VM regardless of duplicate CAs.