RE: Re: FW: I-D ACTION:draft-tsirtsis-dsmip-problem-0 0.txt

 > On Sun, 24 Aug 2003, Soliman Hesham wrote:
 > >  > > => No it doesn't. How does the MN reverse tunnel to the
 > >  > > HA? MIPv4 does not assume reverse tunnelling.
 > >  > 
 > >  > It has worked just fine for me w/ Dynamics 
 > >  > (http://www.cs.hut.fi/Research/Dynamics/), AFAIR.
 > >  > 
 > >  > Are you saying that mobile nodes do not tunnel back to the 
 > >  > HA themselves, 
 > >  > just (optionally) rely on the FA's doing it?
 > > 
 > > => Yes. I don't know what you did with
 > > the Dynamics SW.
 > 
 > RFC3024 and MIPv4 spec at:

=> This is not relevant, this is for topologically
correct addresses, which is not the case in most
MIPv4 deployments. Normally the FA would reverse
tunnel.

 > 
 > So, it seems pretty reasonable to expect many (most?) 

=> More like "none" than "most". Even if they implement
it, the encapsulation is normally done in the FA otherwise
the local domain's ingress filter will drop the packets.

   MN's implement 
 > tunneling back to HA .. if not for any other reason than 
 > being able to 
 > operate in networks where there are no foreign agents.

=> It's not required for operating in such networks. Please
see RFC 3220

 > > 
 > > => what makes the MN reverse tunnel a packet sent to an 
 > > IPv6 address to an IPv4 address?
 > 
 > The enabling of 6to4 pseudo-interface, for example? (i.e., 
 > enabling the 
 > pseudo-interface, resulting in the 2002::/16 route).

=> see above...

 > > => I said WWAN not WLAN. 
 > 
 > I'm not familiar with the term but I assume you mean 
 > something like GSM 
 > networks.

=> Wireless Wide Area Networks, cellular networks. GSM
is one example.

Hesham