Re: [v6ops] DANIR questions draft-shytyi-v6ops-danir-04

On Wed, 13 Nov 2019 at 11:19, Dmytro Shytyi <ietf.dmytro@shytyi.net> wrote:
>
> Hello Mark,
>
> Please find my comments inline.
> ______________
> Dmytro SHYTYI
>
>
> ---- On Tue, 12 Nov 2019 12:12:50 +0100 Mark Smith <markzzzsmith@gmail.com> wrote ----
>
> Missed a few words, touchscreen keyboard!
>
> On Tue, 12 Nov 2019, 22:09 Mark Smith, <markzzzsmith@gmail.com> wrote:
>
>
>
> On Tue, 12 Nov 2019, 20:42 Alexandre Petrescu, <alexandre.petrescu@gmail..com> wrote:
>
> Hello participants to v6ops WG,
>
> With respect to draft-shytyi-v6ops-danir-04, of which I am a co-author,
> I would like to ask:
>
> - do you think it is useful that smartphones run DHCPv6-PD clients on
>    their 4G interfaces?
>
>
> Broadly yes as that would be a conventional IPv6 router WAN behaviour.
>
>
> Great. Thank you for this comment!
>
> Trouble from that perspective with smartphones is how that fits with 3GPPs worldview.
>
>
> - do you think a router behaviour where it runs a Client DHCPv6-PD on
>    one side, derive sub-prefixes and subsequently run SLAAC RA on the
>    other side - has been described somewhere else than in this DANIR
>    draft?
>
>
> Yes, it is described in RFC 7084.
>
>
> In the RFC 7084 the behaviour of IPv6 Customer Edge Router in home or small-office use is described.
> Citation: "ipv6 CE router - is  node intended for home or small-office use that forwards IPv6 packets not explicitly addressed to itself. "
>
> In the Danir draft we consider IoT Router.
>
> Where the IoT devices can be used to enable:
> - numerous IoT applications in farming. (Comment: is home or small-ofice use?)
> - remote health monitoring and emergency notification systems. (Comment: If we consider the human-wearable elements can we say that it is home/small-office use?)
> - the integration of communications, control, and information processing across various transportation systems. (Comment: Transportation system is a home or a small-office?)
> - the procedure to  acquire and analyze data from connected equipment, operational technology, locations and people. It helps to regulate industrial systems. (Comment: is it home/small-office use?)
> ....
>
> Does the the RFC 7084  address the different environments above such as  DANIR daft addresses (DHCPv6_PD and NDP Implementation in IoT Router)?
>

Is this device an IPv6 router or not, per RFC 8200's IPv6 router definition?

What makes an "IoT" device special in this context?

An "IoT" device today is going to be more powerful and have more
memory that a general purpose IPv6 router of 20 years ago (e.g. Cisco
2500).

> I think a variation of DHCPv6
>
>
> a variation of a DHCPv6 relay between the downstream and upstream interfaces would be better because a DHCPv6 relay is DHCPv6 option transparent.
>
>
>
> As far as we mention the DHCP relay we have to consider DHCP clients...
>
> So how we address the next information?
>       Here is a feature request on Android for dhcpv6 client since Jun 3 2012...
>       https://issuetracker.google.com/issues/36949085
> Seems that this feature request was taken into account so far....
>
>

Is Android an "IoT" OS? If you're all about IoT, then I wouldn't think
Android's IPv6 capabilities are relevant.

I don't entirely agree with Google's position however I don't entirely
disagree with it either.

I don't think there is a need for stateful DHCPv6 for addressing (and
it doesn't solve an IPv6 address in use for security purpose some
people think it does), and if a device doesn't need anything more than
DNS settings, then there's no need for a DHCPv6 client at all.

(The use cases I think about regarding using a stateless DHCPv6 client
are supplying NTP time sources, time zone information to a device and
possibly geographical location. Android acquires that information via
other methods).

>
> between the downstream and upstream interfaces would be better because a DHCPv6 relay is DHCPv6 option transparent.
>
> The combination of a client and a server limits the options that can be passed between the upstream and downstream interfaces. To support passing new options, they would have to be upgraded; a DHCPv6 relay wouldn't.
>
> IPv6 CE Device DHCPv6 Option Transparency
>
> https://tools.ietf.org/html/draft-smith-v6ops-ce-dhcpv6-transparency-00
>
>
> Regards,
>
> Mark.
>
>
>
> Yours,
>
> Alex
>
> _______________________________________________
> v6ops mailing list
> v6ops@ietf.org
> https://www.ietf.org/mailman/listinfo/v6ops
>
> _______________________________________________
> v6ops mailing list
> v6ops@ietf.org
> https://www.ietf.org/mailman/listinfo/v6ops
>
>
>