On Sun, Mar 11, 2012 at 5:36 PM, Jeffrey Haas <[email protected]> wrote:

> Brian,
>
> On Mon, Mar 05, 2012 at 07:31:23PM -0500, Brian Dickson wrote:
> > Here is the first of three IDs, concerning the definitions of "route
> leak".
>
>
> : 1.1. Rationale
> :
> :
> :    A route-leak occurs when a prefix is originated by one party,
> :    propagated by other parties, and received by the observer, where the
> :    path used was not intentional end-to-end.  It is a leak if the
> :    receiver did not want the route, from a generic policy perspective.
>
> If the receiver didn't want the route?  While that's perhaps true in many
> cases, the impacted party is usually someone whose route was distributed
> outside the scope where they intended it to be distributed.
>
> -- Jeff
>

Yes, you are correct. We can expand the language however folks would prefer.

(But it is different from saying, "It is a leak _only_ if"... that was
definitely not the intention of the description.)

It should also be noted that, based on the definition(s), it would appear
to be a leak, to every recipient after it became a leak.

And, in most instances, it would also meet other leak-like criteria,
including intended scope.

The important thing is to reduce it to the smallest set of consistently
present conditions -- a basis, if you will -- to make it possible to build
rules into the protocol to stop leaks.

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

Reply via email to