Virtualization Fundamentals

What it is

Virtualization creates a virtual version of something physical, so one real resource can be shared by several workloads.

It saves hardware, power and space, and it makes systems easier to copy, move and back up.

This topic covers three kinds you should know: server virtualization with virtual machines, containers, and VRFs, which virtualize routing on a network device.

How it works

Virtual machines and containers share a server, and VRFs share a router.

Server virtualization

A physical server runs software called a hypervisor. The hypervisor divides the server's CPU, memory, storage and network connections among several virtual machines (VMs).

Each VM runs its own full operating system and behaves like a separate computer, even though they all share one box.

There are two types of hypervisor:

• Type 1 (bare metal): installed directly on the hardware. It is used in data centres because it is efficient and stable.
• Type 2 (hosted): installed as an application on top of a normal operating system. It is common on laptops for testing and learning.

Virtual machines connect to the network through a virtual switch inside the hypervisor. The virtual switch links the VMs to each other and to the physical network cards.

To the physical network, each VM can appear as a separate host with its own MAC and IP address.

Containers

A container packages an application together with the libraries it needs, but it does not include a full operating system. Containers share the kernel of the host operating system, so they are much smaller and start much faster than VMs.

Tools such as Docker create and run containers, and an orchestration system such as Kubernetes manages large numbers of them across many hosts. Containers suit modern application designs made of many small services.

The trade-off is isolation. A VM is separated from others at the hardware level, while containers share a kernel, so a flaw in the kernel can affect all of them.

VRF

A VRF (Virtual Routing and Forwarding) instance is a separate routing table inside one router or Layer 3 switch. Interfaces assigned to a VRF belong to that table only.

One physical router can therefore act as several virtual routers.

This is useful when you must keep traffic apart. A service provider can keep each customer's routes separate, and an enterprise can keep guest, corporate and management traffic from mixing.

Two VRFs can even use the same IP addresses without conflict, because their routing tables never meet.

A simple configuration on a recent Cisco device looks like this:

vrf definition CUSTOMER-A
 address-family ipv4
 exit-address-family
!
interface GigabitEthernet0/1
 vrf forwarding CUSTOMER-A
 ip address 10.1.1.1 255.255.255.0

Adding the interface to the VRF removes any IP address that was already on it, so the address is entered afterwards. Older releases used ip vrf NAME and ip vrf forwarding NAME instead, which support IPv4 only.

To look at a VRF routing table, include the name:

show ip route vrf CUSTOMER-A
show vrf

Without the VRF name, show ip route shows only the global routing table. A ping between two interfaces in different VRFs fails unless you deliberately leak routes between them.