Sorry for chiming in late; I haven't read the draft yet. But the discussion raises an issue I ran into repeatedly.
On 2015-01-19 21:35, Emmanuel Baccelli wrote: > Hi Joe, > > PILC is a great ref that should be cited indeed. Indeed, RFC 3819 is a great basis for teaching, and I have also used it a lot for that. It's usefulness has strangely grown with age, like an old wine. It was written before we embraced RFC 5889, though, so those considerations have to be added. > Another relevant ref is > RFC 5889. One of the recommendations that RFC boils down to: *do not* > use on-link prefixes *at all* on interfaces to multi-hop wireless > networks -- the subject of the draft we discuss now. Yes. We have a whole architecture that is based on off-link membership of hosts to a subnet. What makes this different is indeed: > The reason is: there is no planned structure at link-layer and you > cannot guarantee anything in terms of connectivity over such wireless > interfaces -- which does not however mean that these interfaces are useless! It is important to understand the differences in lifetimes of subnets and links. Subnets are planned and stay for a while; mobility in this model does *not* involve moving between subnets*). Links are almost ephemeral; they come and go with mobility (and other spontaneous changes in radio characteristics). Links are never "complete", indeed. Routers route on paths through links that are all in the same subnet (until, of course, a border router is crossed and we are in a different subnet). Completely different from the old idea of an NBMA, which is a single-link subnet where nodes just aren't able to find each other via multicast. In continuation of the hard-to-root-out "all the world is an Ethernet" thinking, I have also repeatedly run into assumptions not taking into account RFC 5889-style networks, somehow implying that nothing has changed from ATM-style NBMAs. I'm sorry, it's all different. Grüße, Carsten *) Yes, you can *add* traditional inter-subnet mobility as an additional service. That, indeed, is about as well-understood as mobility always has been... _______________________________________________ Int-area mailing list [email protected] https://www.ietf.org/mailman/listinfo/int-area
