Most active commenters
  • jeroenhd(3)

←back to thread

563 points joncfoo | 16 comments | | HN request time: 1.72s | source | bottom
Show context
csdreamer7 ◴[] No.41205035[source]
Can we get .local or .l added for private-use applications too?
replies(5): >>41205064 #>>41205158 #>>41207055 #>>41208856 #>>41210365 #
1. duskwuff ◴[] No.41205158[source]
.local is already reserved for mDNS.
replies(2): >>41205429 #>>41210351 #
2. jeroenhd ◴[] No.41205429[source]
.local is in this weird state where it's _technically_ not reserved, but most PCs in the world already resolve it with special non-DNS software because of the Bonjour/mDNS protocol.

So you end up with the IETF standardising .local, because Apple was already using it, but ICANN never did much with that standardisation.

I doubt ICANN will actually touch .local, but they could. One could imagine a scheme where .local is globally registered to prevent Windows clients (who don't always support mDNS) from resolving .local domains wrong.

replies(4): >>41205459 #>>41205538 #>>41205804 #>>41207138 #
3. arjvik ◴[] No.41205459[source]
Modern windows supports mDNS these days!
replies(1): >>41206329 #
4. candiddevmike ◴[] No.41205538[source]
It's reserved per RFC 6762:

> This document specifies that the DNS top-level domain ".local." is a special domain with special semantics, namely that any fully qualified name ending in ".local.

https://datatracker.ietf.org/doc/html/rfc6762

Applications can/will break if you attempt to use .local outside of mDNS (such as systemd-resolved). Don't get upset when this happens.

Interesting fact: RFC 6762 predates Kubernetes (one of the biggest .local violators), they should really change the default domain...

replies(2): >>41205967 #>>41209086 #
5. throw0101d ◴[] No.41205804[source]
> .local is in this weird state where it's _technically_ not reserved […] I doubt ICANN will actually touch .local, but they could.

It is. See §2.2.1.2.1, "Reserved Names", of ICANN's gTLD Applicant Guidebook:

* https://newgtlds.icann.org/sites/default/files/guidebook-ful...

replies(1): >>41206346 #
6. wlonkly ◴[] No.41205967{3}[source]
But that's an IETF standard, not an ICANN policy. AFAIK there's nothing in place today that would _prevent_ ICANN from granting .local to a registry other than it just being a bad idea.
replies(1): >>41206537 #
7. jeroenhd ◴[] No.41206329{3}[source]
It does! I generally assume mDNS to just be available on every device these days. But I've also seen managed environments where mDNS has been turned off or blocked at the firewall.
replies(1): >>41206566 #
8. jeroenhd ◴[] No.41206346{3}[source]
This document describes the process for requesting gTLDs. Some internal ICANN project could ignore the contents of the guidebook without breaking "the rules". Or they could invent some kind of new TLD system; branded gTLDs didn't exist twenty years ago and I doubt most people would've assumed them to become real, yet blog.google is a real thing that exists.
9. anderskaseorg ◴[] No.41206537{4}[source]
The jurisdictional status of .local and other standards-reserved special use domains is explained by RFC 6761 section 3:

https://datatracker.ietf.org/doc/html/rfc6761#section-3

And ICANN is bound by the IETF/ICANN Memorandum of Understanding Concerning the Technical Work of the IANA, which prevents it from usurping that jurisdiction:

https://www.icann.org/resources/pages/agreements-en

10. dboreham ◴[] No.41206566{4}[source]
mDNS is a broadcast protocol so always "blocked at the firewall ".
replies(2): >>41206985 #>>41210598 #
11. oasisbob ◴[] No.41206985{5}[source]
Multicast too. If you've never needed to manipulate ACLs for multicast traffic, you're not really living.
12. abtinf ◴[] No.41207138[source]
> but they could.

Presumably, ICANN, like any other committee, is not interested in self-castration. Which is what would happen if they challenged Apple.

ICANN could do anything with enough rule changes. And then everyone will ignore them.

13. ◴[] No.41209086{3}[source]
14. mjevans ◴[] No.41210351[source]
Give Apple / mDNS .mdns and let it use THAT instead of .local which should NEVER have been taken from local use in the first place.
replies(1): >>41211753 #
15. ◴[] No.41210598{5}[source]
16. duskwuff ◴[] No.41211753[source]
mDNS (which isn't just an Apple thing) has been using .local for roughly 20 years now. It's a little late to change that.