Re: [homenet] Status of draft-tldm-simple-homenet-naming CFA

[Markus Stenberg <[e-mail]>]

On 10 Aug 2017, at 23.33, STARK, BARBARA H <[e-mail]> wrote:
> 
> With one day left in CFA for draft-tldm-simple-homenet-naming, here is my summary of what I think I've read.
> 
> Exactly 3 people have expressed support for adoption (Daniel [author], Michael R, James). Hmm. That's not a lot.
> 
> Juliusz expressed opposition to adoption, but Ray and Michael said the reasoning for objection was flawed (that Juliusz was setting the bar too high and the procedural objections were not valid in the context of IETF procedures). Ray said the purpose of a CFA is "to get agreement that a document is an appropriate direction for the WG to explore, even if it might require substantial work".
> Ted [author] said he thought it might be reasonable to put the CFA on hold until Daniel did another update.
> Tim C said he thought it was early for adoption (for this and related dnssd drafts).
> 
> I hope I got this summary right. Did I miss anything important?
> Does anyone else have an opinion? Does anyone who has expressed an opinion want to express a new and different opinion?

I find it desirable that a work in this direction goes on. However, there’s details due to which I am not very keen about this particular document (or the related dns-sd documents for that matter, but this is not the forum for those). In order I encountered them during a browse through the document:

- requiring every link on every router to have local DNS forwarder/server seems very broken to me. _one_ in-home DNS server is probably enough.
 ( external dns update could be prevented also by e.g. knowing prefix(es) allocated to homenet, by using ULA, or by judicious firewalling; I prefer ULA but YMMV )

- 3.3
 - it implies that homenet exposes DNS outside home (by default?) and uses instead custom dns server logic to handle .home.arpa from ‘outside’; why not just firewall it and be done with it (or listen only on e.g. ULA prefix)
 - why filter out global IPs?

- 3.5 (PVD madness)
 - WHY? can’t we get just rid of split horizon DNS madness and use _a_ DNS instead of N DNS servers?
 - round-robin = bad (think why happy eyeballs came up for example of why)

I’d much rather see some detail on how selected subset of services can be exposed outside home (including also how related firewalling works), than the PVD stuff, and some of the things seem just misguided from implementation point of view; a set of DNS forwarders/servers seems like overkill if one is implementing N+1 device (which assumes there is ’smarter’ router already in the home, look at the current mesh wireless solutions for example).

Anyway, this is my yearly post quota used for the WG, I’ll be back in 2018 :) Looking forward to using this someday, but given it requires host changes (notably parts of 3.3 and 3.5), I am not holding my breath on that yet.

Cheers,

-Markus