LISP Tutorial

In a traditional IP network, a single IP address has to do two jobs at once: it identifies a host, and it also indicates where that host is located in the network. This works fine as long as a host stays put, but it becomes a problem the moment a host moves — because moving to a different part of the network normally means getting a different address, even though the host’s identity hasn’t actually changed. The Locator/ID Separation Protocol (LISP) was created to split these two roles apart.

Endpoint Identifiers and Routing Locators

LISP introduces two separate kinds of addresses:

• Endpoint Identifiers (EIDs) — assigned to the actual end hosts. An EID stays the same no matter where the host physically moves to.
• Routing Locators (RLOCs) — assigned to the routers that make up the global routing infrastructure. Unlike an EID, the RLOC associated with a given host changes depending on where in the network that host currently is.

A mapping between a host’s EID and its current RLOC ties the two together, and this EID-to-RLOC mapping system is what makes LISP work. By separating identity from location this way, LISP improves routing scalability, makes multihoming more efficient, and allows for better ingress traffic engineering.