Subject: Re: Rolling in IPv6 code into -current
To: None <current-users@NetBSD.ORG>
From: der Mouse <mouse@Collatz.McRCIM.McGill.EDU>
List: current-users
Date: 04/26/1996 09:45:21
> The IETF took the position that the US export policy should not set
> the IETF's technical requirements.  While this does have some value
> as an altruistic statement, I and many others, find it naively
> egotistical.  A fully compliant IPv6 stack must include DES
> encryption and cannot be generally exported at this point in time.

So?  It's pretty US-centric to be concerned about that; why should the
rest of the world - and the IETF in particular - care if the US insists
on shooting itself in the foot over this issue?

Well, that's what I think, though the fact that the IETF chose DES
hints to me that they _are_ being US-centric and not caring about the
rest of the world, instead of the other way around, or they would have
chosen a more secure cryptosystem.  DES is just too weak to be taken
seriously nowadays, and by the time IPv6 ever really happens, if it
ever does, it will be useless, nay, worse than useless (because it will
give naïve users a false sense of security).

> Sigh...  Doesn't matter anyway.  IPv6 is the new ISO of networking.
> It's an overspecified morrass.  ("Hey, let's design a new protocol by
> committee without any real world testing!")

All we ever learn from history is that we never learn anything from
history.  IPv4 was not standardized until it was well burnt in in the
real world; the same is true of (almost?) all successful networking,
protocols and code both.

I don't think IP itself really is lacking anything serious at the
moment; the major problem is routing.  CIDR is addressing that to some
extent, and the technology is improving.  Someone mentioned IP space -
there's _plenty_ of IP space, it's just very badly allocated.  People
are using 8-bit subnets for ten-host ethernets and the like.

That said, it still might be nice to see v6 code, or part of it, in
NetBSD; IPv4's successor won't evolve without people playing with
possibilities, and IPv6 may be a workable place to start playing.

					der Mouse

			    mouse@collatz.mcrcim.mcgill.edu