v4 networks commonly only get one IP for the whole network, and people use NAT with port forwarding to make inbound connections work. With this setup, an attacker only needs to scan the 65536 ports on the router to exhaustively enumerate every single publicly accessible server on your entire network, which is about 3 megabytes of traffic and takes approximately no seconds.

On v6, you don't use NAT and networks are /64. Finding every server requires scanning 65536 ports on all 2^64 IPs, which is about 72 billion petabytes of traffic. There are ways to prune this down somewhat, but however you do it the search space is still far larger.

If you want attackers to not know what's behind your router, you want v6.

> to exhaustively enumerate every single publicly accessible server on your entire network

Enterprise thinking. It's not the publicly accessible servers i worry about, it's the other boxes that shouldn't be publicly accessible...

That's what I meant. On v4, it's trivial to find every server that can be reached from the Internet, whether it was intentional or not. It's not so trivial on v6.

Note that V6 is easier to scan than some people assume. You don't have to scan all 2^128 addresses - you can look at provider address blocks in the registry, and make an assumption (or try it and see) what size block that provider assigns to each server, and then guess the server is ::1 or ::2 in each block. This isn't an exhaustive scan, but you'll find a lot of services this way anyway.

You can also e.g. monitor certificate transparency logs for hostnames. But the difference is that without NAT, knowing about one server on the network doesn't automatically give you the IP for every other accessible server on the same network. You have to actually try host IPs one by one instead of the router kindly filling that part in for you.