Virtual interfaces are network interfaces that are not associated with a physical interface. Physical interfaces have some form of physical element--for example, an RJ-45 male connector on an Ethernet cable. Virtual interfaces exist only in software; there are no physical elements. You identify an individual virtual interface using a numerical ID after the virtual interface name. For example: loopback 0, tunnel 1, and fastethernet 0/0/0.1. The ID is unique per virtual interface type to make the entire name string unique; for example both a loopback 0 interface and a null 0 interface can exist, but two loopback 0 interfaces cannot exist in a single networking device.

Cisco IOS XE software supports four types of virtual interfaces:

• Loopback

• Null

• Subinterface

• Tunnel

A loopback interface can provide a stable interface on which you can assign a Layer 3 address such as an IP or IPX address. This address can be configured as the source address when the networking device needs to send data for protocols such as NetFlow or Cisco Discovery Protocol (CDP) to another device in your network and you want the receiving device to always see the same source IP address from the networking device. This is an issue in networks with multiple equal-cost paths because under normal circumstances the packets that are generated by a networking device use the IP address from the outbound interface as the source address for the packets and because in a network with two or more equal-cost paths from the networking device to the receiving host each packet might use a different outbound interface.

A null interface provides an alternative method of filtering without the overhead involved with using access lists. For example, instead of creating an outbound access list that prevents traffic to a destination network from being transmitted out an interface, you can configure a static route for the destination network that points to the null interface.

Subinterfaces were invented as a method of virtually subdividing a physical interface into two or more interfaces so that the IP routing protocols would see the network connection to each remote networking device as a separate physical interface even though the subinterfaces share a common physical interface. One of the first uses of subinterfaces was to resolve the problem with split horizon on Frame Relay WANs.

The following are several situations in which tunneling (encapsulating traffic in another protocol) is useful:

• To enable multiprotocol local networks over a single-protocol backbone

• To provide workarounds for networks that use protocols that have limited hop counts; for example, RIP version 1, AppleTalk

• To connect discontiguous subnetworks

• To allow virtual private networks across WANs

You can specify a software-only interface called a loopback interface to emulate a physical interface. Loopback interfaces are supported on all platforms. A loopback interface is a virtual interface on a Cisco router that remains up (active) after you issue the no shutdown command until you disable it with the shutdown command. Unlike subinterfaces, loopback interfaces are independent of the state of any physical interface.

The loopback interface can be considered stable because once you enable it, it will remain up until you shut it down. This makes loopback interfaces ideal for assigning Layer 3 addresses such as IP addresses when you want a single address as a reference that is independent of the status of any physical interfaces in the networking device. A good example of this is using the IP address of a loopback interface as the IP address for the domain name system (DNS) host address for the networking device. Before loopback interfaces were available, network administrators had to configure a DNS host entry for every interface on a router that had an IP address assigned to it because they could never be certain which interface IP address might be available at any given time for managing the router. In the sample interface configuration and DNS entries for Router A shown below, you can see that there is a DNS entry for each interface.

GigabitEthernet0 10.10.10.1 255.255.255.0
GigabitEthernet1 10.10.11.1 255.255.255.0
GigabitEthernet2 10.10.12.1 255.255.255.0
GigabitEthernet3 10.10.13.1 255.255.255.0
GigabitEthernet4 10.10.14.1 255.255.255.0
GigabitEthernet5 10.10.15.1 255.255.255.0

RouterA   IN  A  10.10.10.1
          IN  A  10.10.11.1
          IN  A  10.10.12.1
          IN  A  10.10.13.1
          IN  A  10.10.14.1
          IN  A  10.10.15.1

Loopback 172.16.78.1 255.255.255.0
GigabitEthernet0 10.10.10.1 255.255.255.0
GigabitEthernet1 10.10.11.1 255.255.255.0
GigabitEthernet2 10.10.12.1 255.255.255.0
GigabitEthernet3 10.10.13.1 255.255.255.0
GigabitEthernet4 10.10.14.1 255.255.255.0
GigabitEthernet5 10.10.15.1 255.255.255.0

RouterA   IN  A  172.16.78.1

Loopback interfaces provide a stable source interface to ensure that the IP address assigned to the interface is always reachable as long as the IP routing protocols continue to advertise the subnet assigned to the loopback interface. Loopback mode, however, is used to test and diagnose issues with WAN (serial) links such as bit loss or data corruption. The idea is to configure a loop to return the data packets that were received by the interface back out the same interface to the device that originated the traffic. Loopback mode is used to troubleshoot problems by checking that the data packets are returned in the same condition in which they were sent. Errors in the data packets indicate a problem with the WAN infrastructure. Many types of serial interfaces have their own form of loopback command syntax that is entered under interface or controller configuration mode.

The null interface is a virtual network interface that is similar to the loopback interface. Whereas traffic to the loopback interface is directed to the router itself, traffic sent to the null interface is discarded. This interface is always up and can never forward or receive traffic; encapsulation always fails. The null interface functions similarly to the null devices available on most operating systems.

Null interfaces are used as a low-overhead method of discarding unnecessary network traffic. For example, if you do not want your network users to be able to reach certain IP subnets, you can create static IP routes for the subnets that point to the null interface of a networking device. Using the static IP routes takes less CPU time for the networking device than using IP access lists. The static-route configuration is also easier to configure than IP access lists because it is done in global configuration mode instead of in interface configuration mode.

The null interface may not be configured with an address. Traffic can be sent to this interface only by configuring a static route where the next hop is the null interface--represented by Null 0. One example of configuring the next hop to be the null interface is to create a route to an aggregate network that can then be announced through the BGP, or to ensure that traffic to a particular range of addresses is not propagated through the router, perhaps for security purposes.

The router always has a single null interface. By default, a packet sent to the null interface causes the router to respond by sending an Internet Control Message Protocol (ICMP) unreachable message to the source IP address of the packet. You can configure the router either to send these responses or to drop the packets silently.

Subinterfaces are associated with physical interfaces. Subinterfaces are enabled when the physical interface with which they are associated is enabled, and subinterfaces are disabled when the physical interface is shut down.

Note	Subinterfaces can be enabled and shut down independently of the physical port with which they are associated. However, you cannot enable a subinterface of a physical interface that has been shut down.

Subinterfaces are created by subdividing the physical interface into two or more virtual interfaces on which you can assign unique Layer 3 network addresses such as IP subnets. One of the first uses of subinterfaces was to resolve the problem with split horizon on Frame Relay WANs. Split horizon is a behavior associated with IP routing protocols such as RIP in which IP subnets are not advertised back out the same physical interface that they were learned over. Split horizon was implemented to prevent routing loops in IP networks. A routing loop can be created when the networking devices at both ends of a network connection advertise the same IP routes to each other. Split horizon was an issue for Frame Relay multipoint network interfaces--interfaces that connect to two or more remote networking devices over a single physical interface--because the default behavior of many networking devices was to implement split horizon, which means that the networking device did not advertise the IP routes that were learned over an interface back out the interface to other devices that were also reachable via the same physical interface. Subinterfaces were invented as a method of virtually subdividing a physical interface into two or more interfaces so that the IP routing protocols would see the network connection to each remote networking device as a separate physical interface even though the subinterfaces share a common physical interface. Although TCP/IP now disables split horizon limitations by default, protocols such as AppleTalk and IPX are still constrained by split horizon.

Subinterfaces are identified by a prefix that consists of the hardware interface descriptor (IDB) followed by a period and then by a number that is unique for that prefix. The full subinterface number must be unique to the networking device. For example, the first subinterface for GigabitEthernet interface 0/0/0 might be named GigabitEthernet 0/0/0.1 where .1 indicates the subinterface.

Tunneling provides a way to encapsulate arbitrary packets inside a transport protocol. Tunnels are implemented as a virtual interface to provide a simple interface for configuration. The tunnel interface is not tied to specific "passenger" or "transport" protocols, but, rather, it is an architecture that is designed to provide the services necessary to implement any standard point-to-point encapsulation scheme.

There are several ways to implement tunnel interfaces depending on the connectivity that you need to provide. One common use for tunnels is to carry data traffic for a network protocol such as IPX over devices in your network that do not support IPX. For instance, if your network uses IPX in sites at the edge of your network but not in the core of your network, you can connect the IPX sites at the network edges by tunneling IPX in IP over the core of the network.

For more details about the various types of tunneling techniques available using Cisco IOS XE software, see the " Implementing Tunnels " module of the Cisco IOS XE Interface and Hardware Component Configuration Guide.