Re: [lisp] LISP does not involve separate namespaces

[Dino Farinacci <[e-mail]>]

> Short version:  "RLOC" and "EID" are not, and never have been,
>                separate namespaces.

Let me give you an example where EIDs and RLOCs are separate namespaces.

Today, the LISP test network uses 2610:00d0::/32 as a global unique  
EID-prefix. IPv6 sites will have devices assigned from this global  
prefix. This is a PI-prefix because it is not assigned to any service  
provider. This prefix is not injected into the underlying routing  
system, be it the /32 itself or any more specifics.

RLOCs will be assigned out of the 2002::/16 (among possibly other high- 
level prefixes). They can be PA-assigned prefixes. They will be  
assigned to the LISP ETR CE/PE link. They will appear in locator-sets  
of LISP map-cache entries.

So, architecturally, the 2 address spaces are separate and can  
implemented that way. It could be desirable to have an EID address out  
of the 2002::/16 space or a RLOC address out of the 2610:00d0::/32  
space. But it may not be needed with such a large address space.

For IPv4, life is harder because of the vast install base, so the  
clear separation is harder to appreciate. But you could have the same  
address assigned from each namespace.

Just an FYI, in the prototype implementation, we use another VRF  
called "the LISP VRF" which stores solely EID-prefixes so we can  
operate the BGP-ALT in this namespace. And the default VRF has both  
namespaces so you can run underlying BGP at the site as well as having  
EID-subnets in your IGP.

Dino