Re: [v6ops] IPv6-Only Preferred DHCPv4 option

[Brian E Carpenter <[e-mail]>]

As an ex-author of an ex-draft that suggested using IPv6 to tell hosts to avoid IPv4, I'm curious to know whether a draft that suggests using IPv4 to tell hosts to prefer IPv6 will also be accused of being an operational nightmare.

Regards
   Brian

On 05-Dec-19 07:22, Michael Richardson wrote:
> 
> Tommy Pauly <[e-mail]> wrote:
>     > Thanks for posting this draft! I find it to be both a well-written document, and a great solution.
> 
> As an author, thank you.
> I think that this document grew quickly from hallway discussion at IETF106.
> 
>     > As a client OS vendor, I think this would be easy for the client hosts
>     > to implement, and we'd likely see a scenario in which pretty quickly,
>     > most modern mobile devices and laptops would support this, and we'd end
>     > up with using v4 on networks only for more "legacy" devices.
> 
> A question that we had was whether or not hosts in a IPv6-mostly network
> should configure IPv4-LL addresses in order to speak to legacy-v4-only hosts.
> 
> The DHCPv4 option has a payload value that can be returned with information.
> Right now, it's a boolean: if the option exists the network is ready.
> We have considered two things:
> 
> 1) We could put a time at which the host should recheck for v4.  This avoids
>    making mis-configuration of IPv6-mostly networks permanent.  The question
>    is not whether such a recheck timeout should exist (it must); but rather
>    whether we should allow it to be configured in the option, or if we can
>    just suggest some implementation local value.
> 
> 2) Whether or not IPv4-LL (169.254.0.0) addresses should be configured or not.
>    We could make this an option in the payload here as well.
> 
> We wind up with four classes of v4-hosts:
> 
> 1) IPv4-only legacy hosts which do not speak IPv6 at all.
>    (The old printer in the corner, a whole plethora of stupid Web-Connected IoT devices)
>    Some might not even have IPv4-LL addresses though!
> 
> 2) Dual-Stack hosts do not understand this option.
>    They have IPv4-LL and IPv6-LL addresses as well as IPv6 GUAs and ULAs, so
>    they can talk to any host on the network.
> 
> 3) Dual-Stack hosts which have an operational need for IPv4 that is not
>    satisfied by NAT64, and which therefore either do not include this option,
>    or for which the network is aware of their need, and does not answer with
>    the option set.
> 
> 4) Dual-Stack hosts which turn off their IPv4.
>    Do they keep IPv4-LL so that they can talk to (1)?
> 
> [5) IPv6-only capable hosts which never speak IPv4]
> 
> --
> ]               Never tell me the odds!                 | ipv6 mesh networks [
> ]   Michael Richardson, Sandelman Software Works        |    IoT architect   [
> ]     [e-mail]  http://www.sandelman.ca/        |   ruby on rails    [
> 
> 
> 
> --
> Michael Richardson <[e-mail]>, Sandelman Software Works
>  -= IPv6 IoT consulting =-
> 
> 
> 
> 
> _______________________________________________
> v6ops mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/v6ops
>