Re: [dhcwg] [v6ops] Fwd: New Version Notification for draft-link-dhc-v6only-01.txt

[Lorenzo Colitti <[e-mail]>]

On Fri, Dec 13, 2019 at 1:16 AM Ted Lemon <[e-mail]> wrote:

> What you are saying here is effectively that for all eternity, we must
> configure this broken middlebox to believe that some IPv4 address that will
> never be assigned on a link is valid on the link, and configure the DHCP
> server to know what IPv4 subnet is valid on a particular link, even when
> IPv4 is no longer in use on that link.
>

Ted, welcome to the Internet. 'Tis is a place full of middleboxes. Many of
them behave incorrectly. It's great to have you here :-)

But seriously: this sould not be surprising. DHCP has been around for
decades, and yiaddr field is pretty fundamental to the protocol, so you'd
expect there to be some ossification in place. Sure, we can attempt to
ignore it and proceed anyway, but it's a tradeoff. Saying that yiaddr
SHOULD or MUST be 0.0.0.0 creates incompatibility problems and complicates
server implementation. What do we gain by doing so? Why can't we just say
that the server MAY return 0.0.0.0, and if it does so, the client MUST NOT
request 0.0.0.0 but must instead assume that that server has nothing to
offer?