On Wed, Aug 04, 2010 at 02:06:53PM -0400, Rob Austein wrote:
> At Wed, 4 Aug 2010 17:44:24 +0000, Jeffrey Haas wrote:
> > 
> > My concern is that if a cache is out of step with the policy server
> > that is used to populate that cache.
> 
> This description makes no sense in the context of the current protocol
> and implementations.  You've introduced an independent element between
> the RPKI and the rpki-router protocol server (your "policy server"),
> then postulated a failure of that element.  How you handle failures in
> elements you introduce would be between you and your rabbi. :)

Let me attempt to correct my incorrect verbiage.  It's been a while since
I've read the -arch document.

-arch notes that a provider must maintain a Local Cache.

The Local Cache may contain configuration - policy - that directs things
like which TAs to use.  After processing the RPKI objects, the output is
effectively a database of prefixes (and max-lengths) and their origin ASes.

A provider may wish to override validated RPKI data for their own purposes.
While not explicitly a SIDR-driven requirement, this was discussed multiple
times as a requirement during the original RPSEC work.

The resulting output (override or no), is a set of AS:prefix/maxlen tuples
that reflects the provider's routing policy with regard to origin
validation.

This policy must now be distributed to routers for implementation.  The
mechanism for distributing the policy is currently
draft-ymbk-rpki-rtr-protocol.

Implementation/Deployment models:
1. The Local Cache with resulting routing origin policy and the rpki-router
protocol server are in the same box.

In this model, the provider will likely have to deploy more than a single
instance of this model throughout their network.  This will result in each
box hitting the RPKI repositories in order to keep their state fresh.
Presuming the lack of a mechanism to keep each box in lock-step with their
state, this provides a scenario where these servers may become out of synch
with each other.  This may yield inconsistent routing origin policy
within a given network.

2. The Local Cache with resulting routing origin policy is implemented
within one box.  The rpki-router protocol servers are located in multiple
boxes.

In this model, the single Local Cache minimizes the impact on the global
RPKI repository infrastructure.  A mechanism is required to distribute the
resulting origin routing policy to the rpki-router servers (let's call it
scp for simplicity's sake).  The only way for this infrastructure to provide
inconsistent routing origin policy within the network would be for the Local
Cache to be topologically isolated from the rpki-router servers.  This is a
detectable condition at the Local Cache presuming a push model where it is
noted that the push fails.

In both models there is a possibility that the rpki-router server session is
up but the server session is out of sync with the RPKI within a given
network.  This yields inconsistent policy which is usually considered a 
Bad Thing.  If you don't have this perspective, you must be talking to
different providers than I am.

I believe it is possible to minimally update the rpki-router protocol to
allow routers to detect policy skew.

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

Reply via email to