On 3/8/16, 1:47 AM, "j.schoenwael...@jacobs-university.de" <j.schoenwael...@jacobs-university.de> wrote:
>On Tue, Mar 08, 2016 at 01:23:50AM +0000, Acee Lindem (acee) wrote: >> >> The thing about the static route definition for IPv4 and IPv6 is that >>their RIBs will have pretty much the same structure other than >>differences in address type. For other AFs, there may be other >>differences as well. For every augmentation, we’re essentially doubling >>the specification effort. >> > >One benefit of having separate augmentations is that the data types >are tighter. For example, the data model does not allow an IPv6 next >hop for an IPv4 prefix, i.e., you can't mess up address families since >the data model forces a clear separation (and generic validation will >catch that in contrast to having the runtime catching inconsistent >address family data). Actually this is supported on the data center product that I work on (see RFC 5549). However, I agree mixing AFs will not be the norm. > >How much that matters likely can be debated. But somehow it feels like >keeping things separate allows for independent evolution, which may >proof to be very handy in the future. I kind of liked Lada's approach >to maintain the architectural separation in the data model. > >/js > >-- >Juergen Schoenwaelder Jacobs University Bremen gGmbH >Phone: +49 421 200 3587 Campus Ring 1 | 28759 Bremen | Germany >Fax: +49 421 200 3103 <http://www.jacobs-university.de/> _______________________________________________ netmod mailing list netmod@ietf.org https://www.ietf.org/mailman/listinfo/netmod