They thought of that already. One of the differences between IPv6 and IPv4 is that the v6 address space is enormous and stack assumes your interface will have

They thought of that already. One of the differences between IPv6 and IPv4 is that the v6 address space is enormous and stack assumes your interface will have multiple addresses. In particular, on each interface you normally have: * A link-local IPv6 address (non route-able). This is used for neighbor discovery and router solicitation, among other things. * A static-ish cryptographically generated address (often called a secured address). This is used for incoming connections. * One or more temporary addresses with relatively short lifetimes. These are preferred for outgoing connections. The lifetime of these addresses often overlap because any the OS has to maintain the address for any open connections. Note that these addresses can be generated automatically using SLAAC, and you would normally have a set of secured and temporary addresses for each upstream router that is sending router advertisements. You may also have another set of secured and temporary addresses in the IPv6 private range (equivalent to 10.0.0.0), if your router is advertising that (useful for internal services like DNS, where you want to configure the server address in the DHCPv6 config). So it’s not unusual for a given host on a simple home network to have a dozen addresses in 2 or 3 subnets (on one Ethernet interface). The bottom line is that your server should be listening on the secured address and your web browser should be using temporary addresses for outbound connections.