Re: [homenet] Let's make in-home ULA presence a MUST !?

[Gert Doering <[e-mail]>]

Hi,

On Wed, Oct 15, 2014 at 10:34:17AM -0500, Ted Lemon wrote:
> On Oct 15, 2014, at 10:04 AM, Gert Doering <[e-mail]> wrote:
> > I explained my reasoning.  Multiple times.  Here and on other lists.  Again
> > and again.
> 
> When you repeat yourself again and again, people stop listening to you.   

Well, there are some people in the IETF that really don't know how to
listen, but that's not particular to *me*...

Some other people seem to understand what I say, and agree with it, or
sometimes disagree with the particulars, based on practical experience,
or on understanding my points.

> There was a consensus call done on this, and the architecture document 
> contains the results of that consensus.   If you have some additional 
> objection to raise, you should raise it, but I'm very sorry, you do 
> have to show your work.   You can't just make assertions about what you 
> think is true, and expect your opinion to count in the consensus call.   
> We don't vote in the IETF, so opinions just don't make sense in that context.

Huh, what?  Maybe I have problems listening as well.  I fully fail to
see how this paragraph relates to anything I have stated in the thread
- which strayed from the subject of "ULA or not" quite a bit to "what
happens out there in the real world".


> In particular, you appear to be arguing as if ULAs and GUAs are
> treated identically by IPv6 stacks, but they are not.   

This was not my point, and isn't, as it is not relevant.

My point in this discussion is that applications need to handle changing
addresses under their feet, while the application is running, and in
quite a number of cases, while a session is established (to pick
up the "I ssh to my summer house and leave that session running for
months" example - well, add homenet multihoming to the mix, and you 
really want MPTCP, shim6 or something like mosh to handle "oh, and 
I really want my SSH session to stay up even if one of my ISPs goes 
down for a few days").

This is fully independent of the question "do we add ULA to the mix or 
not", as ULAs are irrelevant for communication to partners outside the 
home (as long as nobody puts ULAs into global DNS).

As far as "SHOULD or MUST", I don't have a strong opinion either way,
as local communication is just a special case of global communication - and
if global comms can handle changing addresses, getting that right for
local comms is *so* much easier.  My homenet test setup works nicely
with two ISPs, four partially-meshed HNCP routers, and no ULAs at all, 
but having another address won't make much difference...


> So while I
> agree that there is a real problem making this work for multi-homed
> homenets (a problem, by the way, which homenet has decided to try
> to solve), this is completely orthogonal to the question of whether
> we should use ULAs in the homenet.

Indeed.

Could you remind me what your point was?

Gert Doering
        -- NetMaster
-- 
have you enabled IPv6 on something today...?

SpaceNet AG                        Vorstand: Sebastian v. Bomhard
Joseph-Dollinger-Bogen 14          Aufsichtsratsvors.: A. Grundner-Culemann
D-80807 Muenchen                   HRB: 136055 (AG Muenchen)
Tel: [phone]           USt-IdNr.: DE813185279