Re: [Autoconf] Autoconf addressing model

["Teco Boot" <[e-mail]>]

Hi Charlie,

I never suggested using loopback interfaces are mandatory. I suggest MAY or
SHOULD, or BCP.
For some systems, the overhead or complexity or whatever could be
unacceptable and we should not enforce this (no MUST).


As long as I designed and maintained network infrastructures, I worked with
"/32 management addresses" on loopback interfaces. Still, I have annoying
experiences with routers that select the outgoing interface address as
default source address. Shutting down an interface could have unintended bad
impact on your terminal session. Same for link flapping due to other causes.

For this reason, many routers on the Internet use the loopback interface for
"management". "Management" is the host application on routers. There are
lots of design guidelines for this. My proposal is using the "Internet
lessons learned" in the MANET. Nothing wrong with this, agreed?

As said before, I have also bad experience with applications that bind to
the MANET interface address directly. I am sure I am not the only one.


With MobileIP, we are discussing a host (could be NEMO Router, but then it
acts as host on the visiting link). The MobileIP stack is interested in the
status of the visiting link, but does it propagate this to the applications?
Or use the applications some kind of virtual interface (e.g. a loopback
interface)? Or spoof ifup for interface to home link?


On MANET topology changes reflected in the Internet, I would say: it
depends. With NEMO technology, MR would present its status to the HA. We
could run MANET Routing over an MRHA tunnel, so HA learns prefixes of other
MANET Routers connected to the MR. With redundant HA, things are getting
more and more complex. I would say: reachability shall be reflected in The
Internet, intra-MANET link updates should not. Running BGP in the MRHA
tunnel? Or use the nested NEMO model? See what MEXT will do here, and check
if we can help (e.g. a MANET for NEMO). Or come up with some kind of prefix
delegation for MANETs, where summaries belong to the Border Router and MANET
topology is hidden form the Internet. This is what BRDP provides.


Regards, Teco




|-----Oorspronkelijk bericht-----
|Van: Charles E. Perkins [mailto:[e-mail]]
|Verzonden: dinsdag 3 maart 2009 21:20
|Aan: Teco Boot
|CC: [e-mail]
|Onderwerp: Re: [Autoconf] Autoconf addressing model
|
|
|Hello Teco,
|
|How was it already decided to use loopback addresses?
|
|I hope not, really...
|
|In various implementations of AODV with gateways and
|in other situations, even running Mobile IP to infrastructure
|home agents, we never used loopback.
|
|What would happen if the "autoconf addressing model"
|was the same as the "Internet addressing model"?
|Would that really be so bad?  We could have nice
|things like "multicast", "anycast", unicast, subnets
|and the typical addressing fundamentals that people
|already understand.
|
|Don't we just have to make sure that routing updates
|inside the MANET network don't pester routers in
|the external networks (e.g., Internet)?  I don';t see
|why making that assurance should imply that we have
|to re-architect the whole Internet addressing model.
|
|Regards,
|Charlie P.
|
|
|
|
|Teco Boot wrote:
|> Hi Alex,
|>
|> As far as I know, you are the first person that came up with problems
|using
|> a loopback interface for globally unique addresses / host prefixes
|(/128,
|> /32) for routers. Please provide good argumentation, otherwise we
|follow the
|> already accepted practice in the routing community, also documented in
|> RFC5375 (and others, e.g. RFC3484).
|>
|> Maybe you should post this in v6ops, not in Autoconf.
|>
|> Teco.
|>
|>
|>
|> _______________________________________________
|> Autoconf mailing list
|> [e-mail]
|> https://www.ietf.org/mailman/listinfo/autoconf
|>
|>
|>