Re: [Sidrops] draft-ymbk-sidrops-ov-signal-02.txt

[Randy Bush <[e-mail]>]

hi tim,

(  omg!  someone is alive in this wg!  :)

> I have a question: why is signalling OV (within a trust boundary)
> preferable over doing OV?

within the trust boundary, one tends to want to manage complex services
on fewer devices.  route reflection is a poster child; the clients
have a restricted number of peers and a minimized consistent view.  so
it is what we used in the text.

   Within a routing trust boundary, e.g. an operator's Point of Presence
   (PoP), it may not be desirable or necessary for all routers to
   perform Origin Validation using the Resource Public Key
   Infrastructure (RPKI) per [RFC6811].  A good example is route
   reflectors (see [RFC4456]).

a pop may have a few dozen routers.  no need to run rpki-rtr and full
validation on them all.

> If signalling is desired, then the community seems plausible enough to
> me, albeit that it's not clear to me how this is easier on the router
> than doing OV in the first place.

one can do it with policy, no need for a whole rpki-rtr client and
validation.

randy