wendell on tshoot topologies...

in hunting for the compelling reason to use frame on the 2 tshoot topologies that do not specifically include frame, the best i could come up with is this:

http://www.networkworld.com/community/node/60349

What you get here is the ability to have multiple FR PVCs between each pair of the 5 routers. Having multiple PVCs between a pair of routers allows for layer 3 redundancy between that pair. More importantly, this one physical topology can be configured to create a logical topology that put the routers into a linear sequence (much like the sample TSHOOT topology in the TSHOOT demo), or a full mesh, or anywhere in between.  

So, why FR and not pt-pt? Most of the time, the key ROUTE and TSHOOT layer 3 learning points have no dependencies on the layer 2 technology. And when layer 2 does matter, the majority of the interesting cases use FR. When you really need the routers connected over a LAN or pt-pt, re-cable and re-config as needed.

that's pretty weak and just as i thought... wendell writes "interesting" and i read "frame introduces various finicky behaviors with routing protocols that throw curve balls when you're expecting heat"...

like hyman roth said, "this is the business we've chosen..."

so why not leave ATM in as well?

nobody uses that shit anymore either...

frame should be moved out of ccnp and mpls brought back in...