> Everyone still needs an IPv4 to the outside world for compatibility, and pretty much all the time 24/7, so they wouldn't be saving money at all. Yes, but a m

> Everyone still needs an IPv4 to the outside world for compatibility, and pretty much all the time 24/7, so they wouldn't be saving money at all. Yes, but a more and more services pass to IPv6 you can start to resize your IPv4 infrastructure, and arrive at a point where you will switch it off entirely, as it was done for the transition from analog television to digital one (of course we talk about at least 20 years, but if we don't start it we would never see it). > And why do they need a NAT? My NAT only happens on my personal router, not at the ISP level. It's not something that concerns the ISP at all. They need a NAT because they don't have so many public IP address to give one to each customer. So they introduce a second level of NAT, where a group of customers share the same IPv4 public address. That of course causes a lot of problems, especially for online gaming where it's required to open ports, or in general for applications that use p2p protocols. And so if you want a public IP address dedicated to your connection you have to pay more. I don't know the situation in the US, perhaps your IPS have a ton of public IP addresses they purchased in the past so they can afford to assign to each customer one dedicated address, in my country it's no longer the case with most home connection ISP (and all the mobile ones). > (And if there are legal requirements around recording connections and ports and whatnot, you'll still have to record each connection made under IPv6 as well right? And may I ask where you live that every TCP connection you establish gets logged for 5-10 years? Recording your assigned IP address is one thing -- and for most people's home connections it rarely changes -- but I've never heard of recording every connection.) No you don't. The law requires you be able in case of a crime to identify the customer that originated a particular connection. It means that the police goes to the ISP with the time, IP address and source port of the connection and the IPS has to tell the identity of the customer. The method how this is achieved depends on the technology used. With an IPv4 with public dynamic address, they have to maintain the log of the PPPoE connections, to know at one time the customer that had that address. Not that difficult. With an IPv4 with a NAT, a group of users shares an address, thus the information of only the address is not sufficient to identify the customer that did that connection. So they have to log all the connection opened by the customer, so they an know, from the IP address and the source port of the connection from which customer originated the connection. With IPv6, you don't have any of these problems. With so many addresses you can just statically allocate a /64 block of IP addresses to each customer, and basically don't keep any dynamic record. Of course it's also better for the customer since he has a static IP address if he wants to self host something at home, without using unreliable dynamic DNS services. And nowadays with all the new IoT and domotic stuff having a public address is important (for example just think about accessing security cameras remotely).