Selfhosted

Definitely dual stack if you do. The real benefit of IPv6 is that, supposedly, each of your internal devices can have its own address and be directly accessible, but I don't think anyone actually wants all of their internal network exposed to the internet. My ISP provides IPv6, but only a single /128 address, so everything still goes through NAT.

Setting it up was definitely a learning process - SLAAC vs DHCP; isc's dhcpd uses all different keywords for 6 vs 4, you have to run 6 and 4 in separate processes. It's definitely doable, but I think the main benefit is the knowledge you gain.

There's a bunch of advantages. IPv6 can be useful since your devices can have the same IP both internally and externally. No dealing with port forwarding. No split horizon DNS (where you have different DNS entries for internal vs external). No NAT. No DHCP required for client systems (can just use SLAAC to auto-generate addresses). Much simpler routing. It's a bit faster. Proper QoS.

I used to use Comcast, who actually have very good IPv6 support. They were the first major US ISP to roll out IPv6 to everyone, around 10 years ago. Unfortunately my current ISP doesn't have IPv6, but they're aiming to roll it out this year.

For LAN, no. If you have a router NAT'ting traffic and providing DHCP service there's really no need for ipv6. Almost every ipv6 enabled service provides both 6 and 4 usually and NAT figures it out, and many still provide only 4, meaning you can't just get rid of ipv4 entirely.

If your ISP has modernized and is actually providing an ipv6 address, I suppose there's probably a tiny benefit of being able to go ipv6>ipv6 when routing, bust most all devices nowadays can handle NAT translation from ipv4 to ipv6 and vice versa with no routing penalty. I don't know if there are any ISP's out there who can provide static ipv6 addresses without a NAT router to your entire LAN though.

If you're buying a vps or something ipv6 is easier to get a static address for.

That of course leaves the last good reason: why not? If you're doing homelab hosting stuff why not experiment with ipv6 and fully modernize your network. They suck to type in but it's fun to know your stuff is brand new and using the "best".

Depends on how you define "worth it". Most selfhosting is done not for worth, but for a hobby.

It's good to learn, because it will become more common as time moves on, particularly if you get into the datacenter/cloud/ISP industry. It's less important for the general home user, but it is important to understand how it works and how to use it safely.

Just treating it like IPV4 with more address space is dangerous though. you need to think differently about security and firewalls as it is as if every device has its own dedicated WAN address and could be open to the internet without you knowing.

(Whoops, accidentally hit "Delete" instead of "Edit" and Lemmy doesn't ask for confirmation!! Boo!! I'll try to retype my comment as best I can remember)

I'll buck the trend here and say "Yes, for a home LAN, it's absolutely worth it. In fact for a home LAN it is more important than in a data centre. It is absolutely the bees' knees for home and is worth doing."

All of that depends on how your ISP does things. When I did it, I got a /56, which is sensible and I think fairly common. If your ISP gives you anything smaller than a /64, (a) your ISP is run by doofuses, but (b) it's going to be a pain and might not be worth it. (I now live in literally one of the worst countries in the world for IPv6 adoption, so I can't do it any more)

The big benefit to it is that you can have your servers (if you want them to be) publicly reachable. This means you can use exactly the same address to reach them outside the network as you would inside the network. Just make one AAAA for them and you can get to it from anywhere in the world (except my country).

When I did it, I actually just set up 2 /64s, so a /63 would have been sufficient (but a /56 is nice). Maybe you can think of more creative ways of setting up your networks. Network configuration is a lot of fun (I think).

I had 1 /64 for statically-assigned publicly-reachable servers. Then I had a separate /64 for SLAAC (dynamic) end-user devices, which were not publicly reachable (firewalled to act essentially like a NAT). (Sidenote: if you do go to IPv6 for your home network, look into RFC7217 for privacy reasons. I think it's probably turned on by default for Windows, Android, iOS, etc., these days, but it's worth double-checking)

Dual-Stack is usually no problem, but going IPv6-only is a pain, because a suprising amount of services are v4 only. Even NAT64/DNS64 doesn't help everywhere.

IPv6 is the future so I'd say yes. Dual stack is the way to go. If you can get public address block from your ISP thats great. If not I'd recommend HE tunnel or something similar. Just remember to firewall as ever device is reachable in most configurations.

There aren't many benefits from using IPv6 on LAN, as far as I can tell, unless you need more addresses than are available in the private address ranges.

No, I like living in my nat cocoon so I don't have to worry as much about all the devices on my network. Jk it's turned on, but I don't usually enable it on devices

I'll buck the trend and say "yes, for a home LAN, it is the bees' knees". I don't do it now because my country (and hence my ISP) does not do IPv6, but for most places it's worth doing.

It depends on how your ISP does it. When I did it before, my ISP gave me a /56, which is pretty sensible and I think fairly common. If you get smaller than a /64, (a) your ISP is run by doofuses, but (b) it's going to be a pain and maybe not worth it.

A /56 was much bigger than I needed. I actually only used 2 /64s, so a /63 would have been fine, but network configuration is fun (I think), so maybe you can get creative and think about different ways of allocating your network.

I had 1 /64 for statically-assigned, publicly reachable servers. And then I had a separate /64 for SLAAC (dynamic) allocated personal devices (laptops, phones, etc.) which were not publicly-reachable (firewalled essentially to act like a NAT). (Sidenote, if you are going to use IPv6, I recommend turning on RFC7217 on your devices for privacy reasons. I think these days it's probably turned on by default for Windows, Android, iOS, etc., but it's worth double-checking)

The big benefit to using IPv6 is that all of your home machines can be (if you want them to be) reachable inside your network or outside your network using exactly the same IP address, which means you can just give them a fixed AAAA and access them from anywhere in the world you like. If you're into that sort of thing, of course. It's a lot of fun.

There’s another question: will we ever actually run out of IPv4 addresses, so that cloud providers and ISPs no longer offer them?

I tried converting my internal and external self hosted setup to IPv6 only, like it's the trend nowadays. But halfway through it I couldn't really see the point

Okay, so manu of these answers are just plain wrong. In short, you shouldn't care as the biggest impact will be to network admins. They are the ones who have to configure routing and handle everything else that comes with new addresses. The rest of the world simply doesn't know or notice whether they are using IPv4 or v6. Business as usual.

If the question is whether you should play with it at home. Sure thing if you have the desire to. It's the future and only a matter of time before it becomes a reality. Said network admins and ISPs have been delaying the transition since they are the ones who have to work it out and putting your entire user base behind single IPv4 NAT is simpler than moving everything to IPv6.

From network admin perspective, yes it's worth moving to IPv6 since network topology becomes far simpler with it. Fewer sub-networks, and routing rules to handle those. Less hardware to handle NAT and other stuff. Problem is, they made the bed for themselves and switching to IPv6 becomes harder the more you delay it. Number of users in past 10 years or so has skyrocketed. Easily quadrupled. We use to have home computers with dial-up. Easy enough, assign IP when you connect, release it on disconnect. Then broadband came and everyone is sitting online 100% of the time. Then mobile phones which are also online 100% of the time. Then smart devices, now cars and other devices start having public internet access, etc. As number of users increases, network admins keep adding complexity to their networks to handle them. If you don't have public IP, just do traceroute and see how many internal network hops you have.