In IPv6, what is the thought process of creating subnets on a LAN and VLAN? I know there is no broadcast issue, but what is the thought process on design? Do we have to throw out the old way of thinking when we design IPv6 LANs?
Any and all thoughts on this would be greatly appreciated, as maybe I am over thinking this or just need a mock-up to do a more detailed test and design with IPv6. But all the talk about IPv6 and auto-configuration and letting the router and hosts auto-config is making me think way too much.
Because IPv6 provides a subnet ID as a part of the prefix, as opposed to IPv4 where you have to "steal" it -- so to speak -- from the host bits, there are no variable-length subnet mask (VLSM) considerations. Address design is very simple. It's also important to have the /64 as a maximum prefix length because you need 64 bits of interface ID for host addressing to work correctly (address autoconfiguration).
Some people have been using longer subnets for point-to-point links, but common practice increasingly is to use /64 even for those.
Answer provided by routing expert Jeff Doyle.
This was first published in October 2008