Should you really learn #IPV6 in the #homelab? The curve seems huge! Also is there an upswing? Still what 45 % of the internet only has it.

@train depends on your interests. Compared to other challenges to cope with when running infrastructure, IPv6 is one of the least.

Also, especially in home labs, the real fun starts with #IPv6 since residential connections usually only get a single public legacy IP, but a /56 v6 prefix.

@exponentialverteilt @train

v6 is a great learning tool in a homelab because it allows for interesting routing typologies to be created. It's also amazing if you need to spin up a new prefix inside an existing one you just need to set the routes for the next hop and you're essentially good to go.

No worrying about NAT and the things that break around it when you add more layers of NAT.

@eleix @exponentialverteilt I can see that benefit, but can you expand on the "Routing typologies" that are different that we see in the IPv4 world. It can't just be nat that makes this thing better.

@train @exponentialverteilt

I can certainly try.

In IPv6 I can create a routing topology that can nest almost infinitely without breaking end-to-end connectivity. Note: That is definitely not exclusive to IPv6 and can be done in V4 but would be much more "expensive".

I can take an unused /64, set the next hop address at my core to point to the next "router" in the chain. You would be able to see that router in the traceroute even if it's a PC. Unlike if it was done with NAT.

@eleix @exponentialverteilt I play a network guy on TV, so forgive me.. Couldn't i do that anyway in the same iv4 way. Just add another network somewhere and point the router that knows about that ( using BGP probably) in the core? In ipv6 case, it's just another IP that a router knows about, instead of the routers having to do any translations and keeping that connection together?