As I've been working to adding IPv6 capabilities to NetScanTools Pro. As I have worked with link local addressing I have learn some interesting things. In the IPv4 world if you want to ping an address or connect to a computer on your local network, the software can easily find out the best interface (if you have more than one) to send packets out of to the other computer - all you had to do was supply the IP address. It does this using routing tables and ARP.
In IPv6, it's not quite so easy. In Windows or any other OS, you will often have two or more interfaces capable of talking IPv6. Since all link local addresses begin with FE80:, there is no method for saying this one address is on this segment and this one is on that segment like there is in IPv4 (by using IP address and subnet mask).
Essentially you have to tell the software you are using which interface to use to get to the link local address. If you have one interface to your switch and all the other devices are on the same VLAN, then it's pretty easy. You do this by appending a %# where # is the IPv6 interface number in Windows - also called the Scope ID. In unix derived operating systems like Ubuntu or OS X Lion, you would use %eth0 for the main interface.
How do you know which interface to use? That's where you have to know your network. In my case, this particular XP machine assigns '6' as the Scope ID - you can see this using ipconfig /all. It could be any number, but it's usually a single digit. So to ping an IPv6 address you would enter smoething lik "ping -6 FE80::3CC0:1%6" on the command line (no quotes). Or if you were using NetScanTools Pro v11 Ping Enhanced you would enter FE80::3CC0:1%6 as the target. If you leave off the %6 or change it to another number the software will not know which interface to send the packets out of. In NetScanTools Pro, you will see an error message. The command line ping will tell you the net is unreachable if you use the wrong Scope ID.
Just a few things I've learned about IPv6 link local addressing.