Hacker News Reader

Comment by zamadatix | Hacker News Reader

A good compromise between this (lots of routing work on the internet) and random prefixes all the time (lots of work on the end user networks) is setting DHCP-PD to do consistent PD assignments. Sure, you're space isn't portable between carriers if you move/have redundant internet but you can also rely on a static address in your assigned range working 3 years from now even if your router is offline for a month. ATT fiber does this. I haven't done a survey of who else.

View on HN

ULA is still the right solution here.

ULA would let you maintain your internal lan with custom subnets and DNS even if you switch carriers or use multiple carriers. No need to update your internal DNS servers for ULA.

If you're running a server on your dynamic residential service, you must be using dyndns for ipv4. So do the same with ipv6.

Residential random prefixes is the nature of residential networks, as ISPs don't want to preserve state. With a business grade service, you'd get a static prefix, much like static ipv4. Then the only time you need to update anything is when you switch your ISP and need to update global DNS addresses for your servers.

I guess I'll ask here... How do I learn this? I'm comfortable with IPv4, including DHCP, DNS, VPNs, NAT, whatever. I'd like to be able to set up a v6 network[0] in whatever is the Correct and sane way. Is there a good end-to-end tutorial to set it all up by hand and explain what all these things (RA/DHCPv6/fe80::/ULA/etc.) are and how to use them or why not to use them?

[0] Ideally pure v6 if possible; I think there's some way to encapsulate/NAT v4 traffic out from a pure v6 network so I don't actually need dual stack.

I haven’t seen one. Openwrt seems to do the right thing out of the box, but that’s not that helpful in setting up your own and learning on the way.

Maybe I will write one this weekend and make a hn post.

Hurricane Electric has a pretty decent IPv6 self-study certification program.

https://ipv6.he.net/certification/

For a typical network you usually don't need to do anything. For a geek's network it depends exactly what you're wanting to geek out on. E.g. NAT64 will do what you're looking for at the end (with some caveats depending on what else you're wanting to geek out on) but it's more a carrier focused solution so you don't find as much info about it when reading about consumer tutorials and vice versa.

As for the actual configuration you'll want something specific to your exact network device as learning any of it with a guide for how another device expects the configuration to be entered will be hair pulling.

I wrote one: https://news.ycombinator.com/item?id=40603777

Thank you; that helped:)

Static addresses in your carrier assigned public range are actually very similar to static ULA in terms of what happens with multiple carriers. The same is even somewhat true for moving carriers unless you specifically want to talk to that exact /64 again. That said I'd still recommend cleaning them up in such an event as the main advantage of utilizing the public addresses is things like your DNS servers can be exposed to the internet without the need of NAT. To me that's a lot more the "right solution" than making trade offs for fear of re-iping your devices for the times in your life you change carriers.

Multiple carriers has no (consumer level) solution I've found myself happy with. It's anywhere from "hope all your devices handle multiple public prefixes quickly and without errors" to "NAT and deal with the pain you had from v4 again" to "Use some other way to relay or host your static services which need to be externally accessible" to "go back to dealing with trying to automatically update that DNS server entry on and making sure it always stays consistent"

Reply on news.ycombinator.com