On Wed, Mar 21, 2012 at 8:10 AM, Russ White <[email protected]> wrote:
>
>> A---B---C---D
>>    |       |
>>    +---E---+
>>
>> "A sends an advertisement to B, B sends it to C, but B does not send it
>> to E. Your argument is that BGPSEC prevents D from using the path
>> through E by including in the update a series of signatures."
>
>> I don't see this putting policy in BGP, I don't think anyone has
>> stated that you should send along alternate path info with: "Hey, but
>> don't use this" bits set.
>
> What is the signature's existence in the advertisement from C, but it's
> non-existence in the advertisement from E, supposed to mean to D? Not to
> use the path through E.

but in your example, B never sends the route to E. There is no route,
there is no signature... unless you mistyped and meant that it didn't
sign toward E but still sent a route toward E?

> By adding the signature to the path through C, you've told D not to use
> other paths, even if they really exist. Without the signature, D would

hrm, well... I think you really (backing up and assuming you MEANT
'did not SIGN to E, but sent to E' originally) tell E "Hi, you don't
do security-foo, so you get a plain route", the follow on receipt at C
from E is also not signed (once unsigned, can not resign) C can say:
"Hrm, I prefer Signed over unsigned, except for Customer over Peer" or
some version of that... In any case C has to select the E path as
'better' in some way for it to send the route along to D. At D, again:
"Signed over Unsigned" or whatever their policy is rules the day.

> have to pick the phone up and ask. With the signature, they don't need
> the phone call. If you don't need the phone call to know what B's policy
> is, you've added the policy to the protocol.

again, your example was B never sends anything to E, " B does not send
it to E" (your words, where 'it' refered to the 'advertisement').
So... you can infer all the policy you want at several points in the
picture (as I outlined in my initial response) but... the other bits
about signature are not applicable.

> The point is you've gone beyond the existence of the path here to the
> rightful use of the path --and that is policy.

don't think so.

-chris

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

Reply via email to