Re: [dhcwg] about mip6-hiopt I-D

["Alper Yegin" <alper.yegin at yegin.org>]

Francis,

If you are not objecting to documenting the support for DHCP relay passing
the info to the DHCP server and simply seeking some text that describes the
other possibility of NAS acting as DHCP server, then I'm fine with that. 

But if you are asking us to remove support for DHCP relay options carrying
the info, then that's not fine. 

Alper 

> -----Original Message-----
> From: Francis.Dupont at fdupont.fr [mailto:Francis.Dupont at fdupont.fr]
> Sent: Thursday, April 03, 2008 12:54 AM
> To: Alper Yegin
> Cc: dhcwg at ietf.org
> Subject: Re: [dhcwg] about mip6-hiopt I-D
> 
>  In your previous mail you wrote:
> 
>    As I understand, you are questioning the need for DHCP relay sending
> the
>    AAA-delivered info to the DHCP server.
> 
> => yes, the path followed by the infos is silly IMHO.
> 
>    This is for the case where NAS is co-located with the DHCP relay, and
> there
>    is a separate DHCP server somewhere else in the same "local" network.
> The
>    AAA-delivered info needs to be provided to the DHCP server, so that the
> DHCP
>    server delivers it to the DHCP client. Yes the info needs to zig-zag,
> but
>    that's necessary unless the node acting as DHCP relay assumes the role
> of a
>    DHCP server.
> 
> => I believe you share a misconception with I-D's author: there is no
> such "the server" thing. The DHCP service can be provided by multiple
> servers in parallel (because requests are multicasted at different
> layers). So it is possible and even natural to reply, perhaps partially,
> locally. You give a way to implement it (this is the most efficient)
> but there are others, for instance co-locate a server or simply have
> the relay forwards to a local server too.
> So I still do not like the zig-zag...
> 
> Thanks
> 
> Francis.Dupont at fdupont.fr

_______________________________________________
dhcwg mailing list
dhcwg at ietf.org
https://www.ietf.org/mailman/listinfo/dhcwg