From: changwen (changwen_at_no.spam)
Date: Tue Jan 14 2003 - 15:02:24 PST
The same message as what I just sent but with HTML turned off..
----- Original Message -----
From: "Pekka Savola" <pekkas-at-netcore.fi>
To: "changwen" <changwen-at-attbi.com>
Cc: "Shiao-Li Tsao (Charles)" <sltsao-at-itri.org.tw>; "Liu, Changwen"
<changwen.liu-at-intel.com>; <ngtrans-at-sunroof.eng.sun.com>
Sent: Sunday, March 24, 2002 12:32 AM
Subject: MIPv6 and MIPv4 [Re: (ngtrans) Next steps]
> On Mon, 13 Jan 2003, changwen wrote:
> > > What about IPv6 connectivity when I add an IPv6-only node in IPv4-only
> > > domain? None. And there doesn't (necessarily) need to be that.
> > >
> > > Obtaining the IPv6 connectivity is always the first step.
> >
> > When MNv6 roams into IPv4 domains, three problems: first is the IPv6
> > connectivity, the second is handoff support, and the third is the
> > roaming support in IPv4 domain. As showed in my draft, the first one can
> > be achived via IPv4 tunneling, the second can be achived via MIPv6, and
> > the third can be achived via MIPv4. Also see my answer below.
>
> I'm not sure what you exactly mean by roaming here. I tried to look at
> the draft but couldn't make a head or tail of it (as before when I looked
> it before IETF53), as it contains a lot of assumptions we don't need.
> Clarification (at least a problem statement / applicability up front)
> would be good.
>
> > > > For clarification, in the solution I presented, only MNv6 MUST be
MIPv4
> > > > capable in order to support MNv6 roaming within IPv4 domain.
Howevere,
> > > > the MIPv4 infrastructure in visited IPv4 network is helpful but not
> > > > needed. All MNv6 needs in visited IPv4 domain is a care-of address:
from
> > > > FA if MIPv4 is deployed there or a co-located IPv4 address from e.g.
> > > > DHCP. So there are NO new/additional requirements placed on the
existing
> > > > IPv4 network.
> > >
> > > I'd like to hear why anyone would care about MIPv6 in IPv4-only
domain.
> >
> > The key word here is node mobility. The problem as I stated in my draft
> > and another email is that a dual stack capable MNv6 from an IPv6 site
> > and whose MIPv6 home agent has no IPv4 address roams into IPv4 domain,
> > e.g. with ongoing IPv6 sessions to some other IPv6 nodes, and how we
> > correctly route packets destined to and originated from the MNv6.
> >
> > Is
> > that a real and important problem? Yes.
>
> Disagree.
>
> > Should people care on how to
> > make MNv6 continue functioning in the IPv4 domains? Yes.
>
> Also disagree.
>
> If MIPv6 HA has no IPv4 address, it seems ludicurous to think that its
> MN's would need to worry about IPv4.
>
Now I think you get the question raised in my draft. My draft adds the new
capability to such MNv6s so they can continue functioning in IPv4 domains.
On the question of why such MNv6s need to worry about IPv4, the fundamental
question is whether we should limit the freedom of such MNv6s unreasonbaly
and whether we should support ubiquitous computing/roaming? I think the
answer should be obvious.
Just my two cents.
changwen
> If this problem does not exist with MIPv6 w/ IPv4 address, it seems like a
> non-problem to me.
>
> > > Get IPv6 address with 6to4, then your domain isn't v4-only anymore.
> >
> > When I say a domain is IPv4 only, I mean that the routing fabric of the
> > domain supports IPv4 routing only. However that doesn't exclude any
> > individual node/host running IPv6. I think that is the common definition
> > used in NGTRANS.
>
> Depends on the context, and here the context wasn't clear.
>
> --
> Pekka Savola "Tell me of difficulties surmounted,
> Netcore Oy not those you stumble over and fall"
> Systems. Networks. Security. -- Robert Jordan: A Crown of Swords
>
This archive was generated by hypermail 2.1.7 : Fri Oct 06 2006 - 00:00:32 PDT