Mostly I’m scared I’ll write a firewall rule incorrectly and suddenly expose a bunch of internal infrastructure I thought wasn’t exposed.
Asklemmy
A loosely moderated place to ask open-ended questions
If your post meets the following criteria, it's welcome here!
- Open-ended question
- Not offensive: at this point, we do not have the bandwidth to moderate overtly political discussions. Assume best intent and be excellent to each other.
- Not regarding using or support for Lemmy: context, see the list of support communities and tools for finding communities below
- Not ad nauseam inducing: please make sure it is a question that would be new to most members
- An actual topic of discussion
Looking for support?
Looking for a community?
- Lemmyverse: community search
- sub.rehab: maps old subreddits to fediverse options, marks official as such
- [email protected]: a community for finding communities
~Icon~ ~by~ ~@Double_[email protected]~
Cloud infra engineer here.
Answer: I don’t think about it. Nothing fully supports it, so we pretend it doesn’t exist.
That's exactly my experience with it.
Some certificates are even annoyed by IPv6 and they won't install until i remove any trace of it from the DNS. This should also pretty much be the only occasion I'm forced to deal with IPv6, instead of glancing over it while working on the server configs.
People still use IPv4 because companies are slow to adopt new technologies. They see it as a huge money drain and if there is not a visible or tangible benefit to it then they won't invest in it. IPv6 is definitely a growing technology, it's just taking it's sweet time. For reference, currently the IPv4 has just under a million routes in the global routing table while IPv6 has ~216K routes. About 5 years ago it was something like 100K for IPv6 and not much has changed for IPv4.
I personally do not like the addressing of IPv6. It's not just the length, but now you have to use colons instead of period to separate the octets which leads to extra key strokes since I have to hold shift to type in a colon. It's a minor thing, but when networking is your bread and butter it adds up.
There are also some other concerns with IPv6. Since IPv6 tries to simplify routing by doing things like getting rid of NATing it also opens us up to more remote attacks. It used to be harder to target a specific user or PC that's behind a NATed IP but now everything is out in the open. I'm sure things will get better as more and more people use it and there will be changes made to the protocol however. It's just the natural evolution of technology.
I am very surprised to hear your ISP is not using IPv6. Seems like they're a little behind the times. Unless they just don't offer it to residential customers, which is still a bit behind the times too I guess.
Iv6 doesn't try to simplify routing and remove nat. that's just how things work. Nat is a workaround for ipv4.
Ipv6 is around since 1998. that's not slow to adopt, at that point it is just plain refusal from some because of the costs you mentionend
Ipv6 does simplify routing. It has less headers and therefore less overheard. IPv6 addressed the necessity of NAT by adding an obscene amount of possible IPs. Removing the necessity of NAT also simplifies routing as it's less that the router has to do.
Ipv6 as a concept was drafted in the 90s. It didn't start actually being seriously used until ~2006/7ish.
IPv6 addressed the necessity of NAT by adding an obscene amount of possible IPs
that is correct but doesn't change the fact that nat came afterwards as a workaround und now the ip stack goes back to it's roots without a nat workaround.
It didn’t start actually being seriously used until ~2006/7ish.
true but still nowadays it isn't even slow anymore just refusal
IPv6 has a policy of throwing more address space at stuff to make routing simpler, though.
IPv4 will individually route tiny slices of address space all over the world, IPv6 just assigns a massive chunk of space in the first place and calls it a day.
I think djb was right, over twenty years ago: The IPv6 mess
The IPv6 designers made a fundamental conceptual mistake: they designed the IPv6 address space as an alternative to the IPv4 address space, rather than an extension to the IPv4 address space.
There was an alternative proposal that was backward-compatible with IPv4, but I’ve forgotten the name now.
Oh man, that would have been so great. Think of all the networking stacks that could have just been silently upgraded. Just some letters/numbers appended to the front or back. If you only get x bytes then prepend with zeroes. Adoption would have been mostly transparent.
We turn it off in our office. It doesn’t benefit us.
You could also make the argument that ipv4 through NAT is better for privacy since it obfuscate what, and how many devices are connected to where.
When I was first looking into IPv6, people were talking about how you can self-assign an address by simply wrapping an IPv6 address around your MAC address. But that practice seems to have fallen out of favour, and I'm guessing the reason is, as you say, the whole privacy thing? There's a lot of pushback these days against any tech that makes it easier to fingerprint your connection.
That was so insane - "we need a unique number, let's just use the MAC" - it was like people didn't even think through any of the implications when making ipv6 address schemes.
Similar with the address proposals that ignored the need to minimise the size of core internet routing tables.
IPv6 has privacy addresses, though. Stuff on my network generates a new random address every day and uses that address for outgoing connections, so you can't really track individual devices inside my network.
Just annoyed when I need to specify port when using IPv6. Needs to add square bracket to workaround ambiguity of colon is kinda bad. How can they decide to use colon instead of another special character??
I have IPv6 at home, at work, on my phone, and my hotspot. I have them on my websites and servers. IPv6 is everywhere for me. I use it all the time. Most people do and don't even realize it.
IPv4 still reigns supreme on a LAN, because you're never going to run out of addresses, even if you're running an enterprise company. IPv6 subnets are usually handed out to routers, so DHCPv6 can manage that address space and you don't need to know anything unless you're forwarding ports on IPv6.
For the Internet, just use hostnames. There's literally zero reason to memorize a WAN address when it could be an A/AAAA record.
Company currently uses IPv6! For awhile firewall rules kept biting us as we’d realize something worked in ipv4 but not IPv6 but now I forget it’s even a thing really.
I once paid for a vpc host that was exclusively IPv6 and was shocked how many things broke. I was using it for a discord bot and the discord api didn’t even properly support IPv6 …
IPv6 is now twice as old as IPv4 was when IPv6 was introduced. 20 years ago I worried about needing to support it. Now I don't even think about it at all.
a teammate implemented it because he thought it would be a good resume project. it added more maintenance work to a lot of pieces, forever. there is no measurable benefit to the business
Both my employer and my home ISP use IPv6 since many years now and so does all my own stuff, it's wonderfully convenient to have a globally unique address for everything that I connect to the network.
I try to force everything to use IPv6. It's a huge pain to support IPv4 as a selfhoster. I never had to specify an IP manually, DNS exists for a reason.
IPv6 was "just around the corner" when I was studying 20+ years ago. I kept a tunnel up until the brokers shut down.
I've been hosting some big (partly proprietary) services for work, and we've been IPv6 compatible for a decade.
My ISP finally gave me native IPv6 earlier this year, which gave me the push to make sure my personal hosting does IPv6 as well. Seems like most big players services support it today. It's nice to not have the overhead that CGNAT brings.
IPv6 got a bit of a bad reputation when operating systems defaulted to 6to4 translation but never actually managed to work.
A lot of networks were designed with ipv4 and NAT in mind. There really isn’t a cost benefit to migrate all your DHCP scopes, VLANs, Subnets, and firewall rules to IPv6 and then also migrate 1000’s of endpoints to it.
Much cheaper to just disable ipv6 entirely on the internal network (to prevent attacks using a rogue dhcpv6 server etc) and only use ipv6 on your WAN connections if you have to use it.
I've used IPv6 at home for over 20 years now. Initially via tunnels by hurricane electric and sixxs. But, around 10 years ago, my ISP enabled IPv6 and I've had it running alongside IPv4 since then.
As soon as server providers offered IPv6 I've operated it (including DNS servers, serving the domains over IPv6).
I run 3 NTP servers (one is stratum 1) in ntppool.org, and all three are also on ipv6.
I don't know what's going on elsewhere in the world where they're apparently making it very hard to gain accesss to ipv6.
With NAT existing, I'm not sure there's a significant reason to switch anymore.
Plus the "surprise" privacy and security benefits of just... not having every network connected device directly addressable by anyone else on the global network. The face of the internet and networking in general, plus the security and safety concerns around it, have changed dramatically since v6 was first created.
NAT is just security by obscurity and actually not really security at all. What's protecting you from incoming scans, etc is your network firewall. That firewall works just the same for IPv6. Blocking incoming traffic for your home network is usually the default setting in your ISP issued router anyway.
Working as a network engineer, NAT in a large scale customer environment can quickly devolve into a clusterfuck. Many times we had week long reachability issues due to intermediate ISPs NATing unexpectedly.
My nemesis is GCNAT, which adds another layer of NAT because some ISPs don't have enough public IP space for all their customers to go around.
I have a customer where their ISP just assigned one of their locations public IPv4 addresses. Neither the customer, nor the ISP owned that address space. Their logic was that this address space is registered on a different continent, so it's basically fair game to use it themselves. Granted, they only route it internally for a MPLS network, but still...
What I'm getting at is that NAT increases complexity and breaks properly routed end to end connections. Everyone kinda fucks up with NAT, especially ISPs (in my opinion anyway).
I can really recommend the IPv6 study material from the major internet registries (took the v6 courses from RIPE NCC myself).
IPv6 is so much simpler for subnetting, writing firewall rules,... IMO the addresses just look kinda clunky.
I want to love IPv6 but it's unfortunately still basically impossible to get good proper IPv6 in the first place.
At home I'm stuck with fairly broken 6rd that can't be hardware accelerated by my router and the MTU is like 1200 which is like 20% bandwidth overhead just for headers on the packets.
On the server side, OVH does have IPv6 but it's not routed, so the host have to pretend to have all the IPv6 addresses and the OVH routers will only accept like 8 of them in use before its NDP table is full, so assigning an IPv6 to every Docker container fails miserably.
IPv6's main problem is ISPs are so invested in NAT and IPv4 infrastructure they just won't support IPv6. Microsoft, Google and Apple need to team together and start requiring functional IPv6 to create user demand, because otherwise most users don't know about CGNAT and don't care. Everything needs to complain about bad IPv6 connectivity so users complain to ISPs and pressure them into fixing it.
I enjoyed getting the IPv6 certification from Hurricane Electric. Everyone should learn about it! https://ipv6.he.net/
Widespread IPv6 adoption is right there with the year of the Linux desktop. It's a good idea, it's always Coming Soon™ and it's probably never going to actually happen. People are stubborn and thanks to things like NAT and CGNAT, the main reason to switch is gone. Sure, address exhaustion may still happen. And not having to fiddle with things like NAT (and fuck CGNAT) would be nice. But, until the cost of keeping IPv4 far outweighs the cost of everything running IPv6 (despite nearly everything doing it now), IPv4 will just keep shambling on, like a zombie in a bad horror flick.
Another thing that makes no sense is if my ISP provided prefix changes -which it will- this affects the IP addressing on my local network. Ain't noboby got time for that if you're managing a company or having anything other than a flat home network with every device equal.
IPv6 is just people shouting NAT BAD, but frankly having separate address ranges inside and outside a house is a feature. A really really useful feature. Having every device have a public IP6 address I'd an anti-featute.
if my ISP provided prefix changes... affects the IP addressing on my local network.
IPv6 is just people shouting NAT BAD... Having every device have a public IP6 address I'd an anti-featute.
If you're working in IT then you should find a new career.
We mainly use ipv4, but recent laws that all public sector websites are to use IPv6, we have had to update our stack.
Now we can do IPv6 public endpoints with ipv4 backends.
Have been using it since late 90s, stopped using it with the shutdown of SixXs as there still were no viable native options in pretty all my infra locations. Recently started using it again as I finally have an ISP providing proper v6.
As an email guy, I would love IPv6, but it just isn’t gonna happen (for me).