> nope, it's bits that you can choose to use or not.

In other words, you've added another set of "wiggle words" to your
definition. "It's not policy because it's not signing something saying
you shouldn't do x, and because anyone can ignore the signature anyway."

Sorry, I don't buy it.

You are adding bits that allow (even demand) a receiver to discern the
intent of the transmitter about the usability of a specific path. I
don't care how many wiggle words you put around it --it's still policy.

Now if you're going to dive into policy, then dive into policy. There's
nothing wrong with diving into policy, so long as you actually address
the issue. That means determining which policies should be protected,
and what the tradeoffs are for protecting each one.

> I'm really not trying to sneak anything.  we seem to be at an impasse.

I've seen more than a small number of presentations on BGPSEC. I've read
more than a small number of drafts. I've had more than a small number of
private conversations about BGPSEC/S-BGP. In _every_ conversation I've
ever had, in _every_ presentation I've ever seen, the one singular
reason given for signing the BGP update is this: So a remote AS can
determine whether or not some previous AS _intended_ to allow
reachability through a particular peer for a particular prefix.

Perhaps the way to resolve the impasse is to accept that what you are
trying to do goes beyond simple "bits on the wire" security, bring the
policy problems out into the open, and have a discussion about them.

Or to come to the realization that you simply can't secure any routing
protocol by simply "signing the bits on the wire," because many of the
semantics everyone is so concerned about are actually not in the "bits
on the wire" to be signed in the first place.

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

Reply via email to