Re: dynamic v4 addresses over v4-in-v6 tunnel [RE: DSTM]

[Francis Dupont <Francis.Dupont@enst-bretagne.fr>]

 In your previous mail you wrote:

   > 3.  The use of IPv4 addresses for legacy communications should be
   > temporary IPv4 addresses for a specific time.
   
   I'm not sure I personally buy the other requirements, but for the 
   argument's sake..
   
   What kind of requirement is this?  How is this different for a 
   site currently deploying v4?
   
=> one advantage of DSTM is you no more need to manage an IPv4
infrastructure (no IPv4 routers, etc). And of course all features
provided in your IPv6 are inherited.

   In other words: if we need IPv4-in-IPv6 tunneling, let's discuss that.  
   If we need something more than the existing IPv4 mechanisms for making
   the v4 addresses used over such tunnels more dynamic (e.g., if DHCPv4
   with short leases is not enough), let's discuss that separately as
   well.  But let's not get these two features mixed up, having to
   sacrifice a baby to get the bathwater :)
   
=> the whole is more than the addition of all parts.

   I have the impression that a lot of people have "IPv6-in-IPv4 == DSTM"  

=> IPv4-over-IPv6 == DSTM ?

   in their head.  I want to break that assumption, as DSTM seems to
   offer a lot more than that, because specifying a solution without
   clearly defining its components seems unwise.
   
=> I agree. To come back to the real point IMHO this is like to
make the same for DHCPv4 itself: DHCPv4 can allocate temporary addresses
for a specific time but it can do more (BTW not only DSTM can use the
same mechanisms than DHCPv4 but my old implementation of DSTM exactly
reuse the code of a well known open source DHCPv4 server :-).

Thanks

Francis.Dupont@enst-bretagne.fr