[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [69ATTENDEES] DHCP

On Sep 13, 2007, at 14:11 , Shane Kerr wrote:


(This is a bit silly, because everything other than "IPv6 has more
addresses" has been added to IPv4 - of course!)



Why is this silly? And _stateless_ autoconfiguration with stable
addresses can't be added to IPv4 because the address space isn't
large enough.



Yet Microsoft and Apple both do IPv4 autoconfiguration. Not perfect,
but it works.
If you mean "you get a different address each time" and "only works in the absense of connectivity", then, sure. But lots of things work much better if connectivity to the internet at large is not a requirement. 



Because of this, they resisted DHCPv6, and they continue to oppose
DHCPv6.



Hm, how exactly did the IPv6 designers in ~ 1995 resist DHCPv6 that
wasn't published until 2003?



The first DHCPv6 draft was published in February 1995, and the
resistance kept it from being published for 8 years. There were *28*
drafts before it was finally moved to RFC status!


Hm, I didn't know that.
And they still didn't get the stateful/stateless thing right (= the client must decide which type to use even though the server is in a better position to know what's appropriate). 

So many cycles is never a good thing.


I suppose I am "in the DHCP camp", but my take on it is not to
oppose network designs without DHCP, but rather that we should not
attempt to duplicate the functionality of DHCP in other protocols.
By that logic, we should all be running RIP because OSPF is a duplication of functionality. 


So we're free to NOT use DHCP as long as we forego the functionality
that DHCP provides? I'm sorry, that's not a reasonable position.



Hey, it's your network, do what you want. If you want to configure
manually, go ahead. If you want to use a home-grown protocol, go
ahead. But if you want an IETF-approved way to do it, then you must
play nice with the IETF.
The IETF is not in the business of telling people how to run their networks. And, fortunately, the IETF is not the same as the dhc wg. 


The opposition to duplicating DHCP functionality in other mechanisms
is the same as if you proposed an alternative to *any* existing
protocol. You should explain why the benefits your new protocol brings
cannot simply be added to the existing technology.
Your assumption is that it's reasonable for me to implement DHCPv6 in my network to address a given need. That assumption is incorrect.