Re: [OSPF] OSPF Topology Transparent Zone (TTZ) Next Steps

[William McCall <[e-mail]>]

Seems like, to me, the net accomplishment is to avoid a L2VPN layer 
sitting where the internal TTZ nodes are, based on the scenarios 
proposed in these drafts. Looks like it would also allow a lot more 
flexibility in adding nodes to the TTZ vs. trying to do L2VPNs, 
particularly with regard to metric calculation (but also future growth 
if an internal node is made edge).

 From draft-chen-ospf-ttz-05:
>   A TTZ MUST hide the information inside the TTZ from the outside.  It
>     MUST NOT directly distribute any internal information about the TTZ
>     to a router outside of the TTZ.

One thing I'm not so happy about in the drafts is that they seem to 
imply that ANYTHING from the internal TTZ nodes is not sent to the 
TTZ-external nodes from an OSPF perspective. IMO, I'd want the option to 
advertise a loopback (or whatever) from the internal TTZ nodes out to 
the TTZ-external nodes so I can reach them for management, etc. I guess 
this could be done with BGP, but that seems overkill for what I'm 
looking for.

On 07/06/2013 03:28 PM, Russ White wrote:
> In general, I think this is useful. I can think of specific situations 
> where it would be nice/would have been nice to have. 
My opinion would be to adopt it as a WG draft, but with a bit of 
scrutiny as to the use cases. Russ got it right when he said there are 
"specific situations", but I guess the question would be around whether 
the situations are prolific enough to warrant work on this. I vote 
"probably", but smarter people than me know better.

--wm