Map and Encapsulate
LISP enables what is broadly referred to in the networking industry as an overlay. Figure 2-1 illustrates the main elements of an overlay service. In it, two planes of functionality enable an overlay network:
Virtual network in the overlay plane
Transport network in the underlay plane
Figure 2-1 Functional Components of a Network Overlay
The underlay plane is a traditional network, which provides connectivity between network devices (routers and switches) but isn’t aware of the endpoints that attach to the network edges. Multipathing and resiliency are optimized in the underlay network with well-understood traditional routing methods. The underlay handles routing only between RLOC addresses.
The overlay plane is a virtual network service that is delivered over the top of the underlay network. The overlay functionality is enabled at the edges of the network only. Traffic between hosts is tunneled between network edge devices across the underlying core network. To determine where to tunnel the traffic to, the edge devices need to obtain the information regarding which edge device a particular host destination may be connected to. This process of mapping identity to location to encapsulate traffic to the destination’s location is often referred to as map and encapsulate.
The LISP functionality is enabled mainly at the edges of the network. From the LISP perspective, the edge devices where LISP is enabled are referred to as tunnel routers. Because the role of the tunnel router is directional, ingress tunnel routers (ITRs) and egress tunnel routers (ETRs) are used, referring to the ingress to the LISP overlay and egress from the LISP overlay, respectively. It is common to see the general role of an edge device referred to as an xTR when directionality is not relevant. The roles and responsibilities of the different types of xTRs are defined in more detail later in the “LISP Roles” section of this chapter, but it is worth noting at this point that requesting a mapping and encapsulating the traffic are ITR functions.
An ITR uses tunnels to encapsulate EID traffic and transport it over the RLOC underlay. From this perspective, there is an inner header in the EID space and an outer header that uses RLOC addresses. Thus, an ITR can encapsulate traffic for any type of EID address family into tunnels using any type of RLOC address family. For example, ITRs may encapsulate traffic for IPv4 EIDs using an IPv6 outer header, or ITRs may encapsulate traffic for MAC EIDs using an IPv4 outer header. Any combination is possible, in theory, and does not affect the way in which the LISP control plane operates.