> Making the spec PS fits the envisaged evolution of the LISP "experiment".
> The downside here might be that this is also recognition that LISP does
> not fix the _global_ routing scalability issue (presuming of course that
> the issue remains) in any short term exercise due to the way operators
> actually choose to run BGP. Is the global internet ready and willing to
> separate the locator and identifier in operations yet? I don't think so.

I think it does fix the problem or at least helps the problem Terry. The real 
question, as you state, is the Internet ready to use this solution.

> I would certainly say that it makes intra-AS routing systems leaner, as a
> very positive outcome.

And netflow caches smaller and EIDs in the underlay possibly private and more 
secure. All good things for USERS.

> I think the scrutiny came through multiple different lenses. Not just
> global routing scalability. I also doubt that the scrutiny will diminish.
> But I do like the idea of calling it what it really is. It's an overlay
> with attractive properties.

Agree.

> Take care to not leave the door _wide_ open. I look at LISP as a very
> (very!) well reviewed specification that currently is not overweight with
> unused/unnecessary features. I have a concern that having all and sundry
> adding in their pet feature may cause that attribute to suffer. Perhaps it
> would be sane to pick the 4-5 priority items and work on those, nail them
> and move forward that way.

Agree 100%, but Prague will be used to brainstorm. Once we do that, we need to 
DECIDE on those priority items. Hope you will be there to contribute. You have 
a lot of experience with LISP deployemnt and can bring a lot of value to the 
discussion.

Thanks,
Dino

> 
> Cheers
> Terry
> 
> 
> 
> 

_______________________________________________
lisp mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lisp

Reply via email to