Comments on: Why You’ll Want LISP Routing – Part 2

By: Omar Sultan

Omar Sultan — Thu, 08 Nov 2012 19:26:01 +0000

You are correct–thanks for the catch.

Omar

0 likes

By: Matt S.

Matt S. — Thu, 08 Nov 2012 15:08:21 +0000

Is the addressing depicted in the drawing “LISP-2-4″ correct? Shouldn’t the red SRC/DST pair be 10.10.1.1/10.100.19.21 to illustrate that the original packet is being encapsulated by the Ingress Tunnel Router?

0 likes

By: Dino Farinacci

Dino Farinacci — Wed, 07 Nov 2012 02:41:22 +0000

Yes, the first set of packets are dropped until the map-cache entry is populated in the ITR. However, an ITR could encapsulate to an RTR (a Re-encapsulating Tunnel Router) where it has map-caches populated by many of sources so the chances for map-cache misses are minimized.

Dino Farinacci
Cisco

0 likes

By: Matt

Matt — Tue, 06 Nov 2012 17:58:59 +0000

Great explanation!

What about an uncached scenario where the RLOC is not known before the first packet of a “flow” reaches the xTR. Those packets are dropped in my test environment until the xTR has enough information on how to forward LISP traffic (by doing a MAP request).

Is this behaviour intentional?

IMO it has a rather strong impact on user experience (page load delays with HTTP etc.)

What do you think about it?

0 likes