Who benefits from this? Even though Let’s Encrypt stresses that most site operators will do fine sticking with ordinary domain certificates, there are still scenarios where a numeric identifier is the only practical choice:

Infrastructure services such as DNS-over-HTTPS (DoH) – where clients may pin a literal IP address for performance or censorship-evasion reasons.
IoT and home-lab devices – think network-attached storage boxes, for example, living behind static WAN addresses.
Ephemeral cloud workloads – short-lived back-end servers that spin up with public IPs faster than DNS records can propagate.

all 43 comments

sorted by: hot top controversial new old
[–] 51 points 1 year ago (16 children)

That's kind of awesome! I have a bunch of home lab stuff, but have been putting off buying a domain (I was a broke college student when I started my lab and half the point was avoiding recurring costs- plus I already run the DNS, as far as the WAN is concerned, I have whatever domain I want). My loose plan was to stand up a certificate authority and push the root public key out with active directory, but being able to certify things against Let's Encrypt might make things significantly easier.

  • source
  • hideshow 16 child comments
  • [–] 17 points 1 year ago (3 children)

    FYI you can get a numeric xyz domain for 1$ a year

  • source
  • parent
  • hideshow 3 child comments
  • [–] 7 points 1 year ago (9 children)

    I use a domain, but for homelab I eventually switched to my own internal CA.

    Instead of having to do service.domain.tld it's nice to do service.lan.

  • source
  • parent
  • hideshow 9 child comments
  • [–] 3 points 1 year ago (8 children)

    Any good instructions you would recommend for doing this?

  • source
  • parent
  • hideshow 8 child comments
  • [–] 3 points 1 year ago

    I just use openssl"s built in management. I have scripts that set it up and generate a .lan domain, and instructions for adding it to clients. I could make a repo and writeup if you would like?

    As the other commenter pointed out, .lan is not officially sanctioned for local use, but it is not used publicly and is a common choice. However you could use whatever you want.

  • source
  • parent
  • [–] 3 points 1 year ago (6 children)

    use the official home.arpa as specified in RFC 8375

  • source
  • parent
  • hideshow 6 child comments
  • [–] 4 points 1 year ago (5 children)

    No thanks. I get some people agreed to this, but I'm going to continue to use .lan, like so many others. If they ever register .lan for public use, there will be a lot of people pissed off.

    IMO, the only reason not to assign a top-level domain in the RFC is so that some company can make money on it. The authors were from Cisco and Nominum, a DNS company purchased by Akamai, but that doesnt appear to be the reason why. .home and .homenet were proposed, but this is from the mailing list:

    1. we cannot be sure that using .home is consistent with the existing (ab)use
    2. ICANN is in receipt of about a dozen applications for ".home", and some of those applicants no doubt have deeper pockets than the IETF does should they decide to litigate

    https://mailarchive.ietf.org/arch/msg/homenet/PWl6CANKKAeeMs1kgBP5YPtiCWg/

    So, corporate fear.

  • source
  • parent
  • hideshow 5 child comments
  • [–] 1 point 1 year ago (4 children)

    But home.arpa’s top-level domain is .arpa?

  • source
  • parent
  • hideshow 4 child comments
  • [–] 1 point 1 year ago (3 children)

    I'm not sure I follow the question. All of the TLD *.arpa is not reserved for private use, only *.home.arpa. So all your internal services are required to be a sub domain.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 1 point 1 year ago (2 children)

    Sounds like you followed.

    Now that I’m moving goalposts, why not use .home.arpa subdomains?

  • source
  • parent
  • hideshow 2 child comments
  • [–] 14 points 1 year ago (6 children)

    F I N A L L Y

    Now tell me it supports IPv6 and I'll be the happiest man alive

  • source
  • hideshow 6 child comments
  • [–] 2 points 1 year ago* (5 children)

    Maybe I'm not understanding it but I can't see what I would use this for due to the 6 day issue period. Bringing a NAS up to copy data for a couple days is the only real use case I find for home users.

    Because even if you pay for a static external IP from your ISP, this doesn't support using such for longer than that period right?

  • source
  • parent
  • hideshow 5 child comments
  • [–] 1 point 1 year ago (2 children)

    Can't it automatically be renewed?

  • source
  • parent
  • hideshow 2 child comments
  • [–] 8 points 1 year ago (1 child)

    Let's Encrypt is meant yo be used with automated certificate renewal using the ACME protocol. There are many clients for this. Both standalone and built into e.g. Caddy, Traefik and other software that does SSL termination.

    So this specific concern doesn't really make sense. But that doesn't mean I really see a use case for it either, since it usually makes more sense to access resources via a host name.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 7 points 1 year ago*

    This could go a long way towards fighting online censorship. One less issue when an authoritarian overreach gets your domain seized. Pretty awesome.

  • source
  • [–] 6 points 1 year ago (1 child)

    Would this work with a public dynamic DNS?

  • source
  • hideshow 1 child comment
  • [–] 16 points 1 year ago (1 child)

    Its like self signed certs with the convience of a third party

  • source
  • hideshow 1 child comment
  • [–] 3 points 1 year ago

    Maybe kinda, but it's also a third party whose certificates are almost if not entirely universally trusted. Self-signed certs cause software to complain unless you also spread a root certificate to be trusted to any machine that might use one of your self-signed certs.

  • source
  • parent
  • [–] 2 points 1 year ago (5 children)

    I never understood why we don't use IP certificates to encrypt the domain with SNI.

  • source
  • hideshow 5 child comments
  • [–] 3 points 1 year ago (4 children)
  • [–] 9 points 1 year ago (2 children)

    In much simpler terms:

    Think of an IP address like a street address. 192 My Street.

    There might be multiple businesses at one street address. In real life we address them with things like 1/192 My Street and 2/192 My Street, but there's no direct parallel to that in computer networks. Instead, what we do is more like directing your letter to say "Business A c/o 192 My Street". That's what SNI does.

    Because we have to write all of that on the outside of the envelope, everyone gets to see that we're communicating with Business A. But what if one of the businesses at 192 My Street is highly sensitive and we'd rather people didn't know we were communicating with them? @bjoern_tantau@swg-empire.de's proposal is basically like if you put the "Business A" part inside the envelope, so the mailman (and anyone who sees the letter on the way) only see that it's going to 192 My Street. Then the front room at that address could open the envelope and see that the ultimate destination is Business A, and pass it along to them.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 5 points 1 year ago

    Currently before establishing an encrypted connection to a webserver the domain is sent to the webserver unencrypted so that the server can choose the appropriate certificate to use for encryption. That is called SNI, Server Name Indication.

    Of course that's a privacy risk. There are finally protocols to fix this but they aren't very widespread and depend on DNS over HTTPS.

    I think issuing certificates based on the IP and sending the domain name encrypted based on that certificate could have fixed this issue ages ago.

  • source
  • parent
  • [–] 2 points 1 year ago (1 child)

    I would read layman discussions about why this (by context?) is good.

  • source
  • hideshow 1 child comment
  • [–] 5 points 1 year ago

    Domains need to be registered annually and DNS servers are needed to route traffic to them. But using an IP directly, you don't need to worry about domain registration issues that can brick your systems, and you don't have to worry about DNS providers knowing about your traffic (or maintaining your own private dns).

    If it's not a user trying in a memorable domain, an IP serves much better.

  • source
  • parent
  • [–] 0 points 1 year ago (4 children)

    Couldn't this prove very troublesome in combination with carrier grade nat?

  • source
  • hideshow 4 child comments
  • [–] 9 points 1 year ago* (last edited 1 year ago) (2 children)

    I don't see how? Normal HTTP/TLS validation would still apply so you'd need port forwarding. You can't host anything on the CGNAT IP so you can't pass validation and they won't issue you a cert.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 1 point 1 year ago (1 child)

    You can totally host something on carrier-grade NAT using techniques like NAT hole punching.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 6 points 1 year ago

    You don't get control of the incoming port that way. For LetsEncrypt to issue a certificate primarily intended for HTTPS, they will check that the HTTP server on that IP is owned by the requesting party. That has to live on port 80, which you can't forward on CGNAT.

  • source
  • parent
  • [–] 2 points 1 year ago

    They will require the requester to prove they control the standard http(s) ports, which isn't possible with any nat.

    It won't work for such users, but also wouldn't enable any sort of false claims over a shared IP.

  • source
  • parent