On Aug 9, 2011, at 9:42 PM, George Michaelson wrote:
> 
> Also, "newly designed" seems a bit strong. This is BGP + signing chains isn't 
> it? Its not re-entering from the top clean state.

A BGPSEC-enabled routing system looks a lot different than BGP...

But let's say "new functionality" for periodic v. triggered updates, you ok 
with that?  :-)

> I am still struggling to understand how Path prepend is going to work. What I 
> heard suggests its going to have to be administratively constrained to be 
> sign-able. At the edge its more in the hands of the origin AS but beyond that 
> where does the permission to play with the path come from?

That's orthogonal, perhaps you should have a look at Doug's slides and mailing 
list discussion related to pCNT, which has it's own set of issues.

> I guess I live in a margin where they are  research TOOL and you sometimes 
> remove TOOLS. If they were added for another purpose, what I get from them 
> (which is not much btw, but they get talked about in my hearing) is not the 
> core motivation.

Again, adding periodic updates to BGP is far from a trivial change from where 
we are today, and simply to "reduce the vulnerability window for replay 
attacks" I'm not convinced we can get to any frequency where such a fundamental 
architectural element is worth the investment.

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

Reply via email to