Just wondering if anybody with IPv6 knowledge can put me straight. I am currently studying for my CCNP (BSCI) and am just going over the IPv6 stuff again.
Now I understand that for the CCNP I will likely just need to know the basic theory of IPv6, however I cannot get my head around the Link-Local and Site-Local addresses.
Routing TCP/IP Vol 1 (Jeff Doyle) 2nd edition, page 54, has a table showing the high order bits of various IPv6 addresses and lists the following.
Link-Local = 1111 1110 10 which equates to FE80::/10
Site-Local = 1111 1110 11 which equates to FEC0::/10
Now the part I can not get my head around is this.
The first 10 high order bits equate to 2 1/2 (10 bits) nibbles, but for the Link-Local address to be FE80 would require 4 nibbles (16 bits) or
1111 1110 1000 0000. Like wise with the Site-Local address of FEC0 which would be 1111 1110 1100 0000. This would mean that the prefix should be /16 if remaining 6 bits can not be turned on.
So I am trying to understand what happens to the remaining 6 bits in these Local IPv6 addresses or to the IPv6 addresses from FE81 to FEBF and from FEC1 to FECF? Are they just ignored?
If so it would appear that nothing was learned from what happened with the initial wasteful assignment of IPv4 addresses!!
Can anybody put me out of my misery please :)
It's a little trickier than that. I never knew why they listed them as /10's because you're right, it implies things can be set however you want, but they can't.
Link local is specifically:
FE80:0000:0000:0000:: (eui-64 or defined address)
Site local is specifically:
FEC0:0000:0000:(subnet):: (eui-64 or defined address)
To further muck things up, the site local addressing has been deprecated in the real world. There is a new set called ULA (unique local address) out there, but it won't be covered in the CCIP exams yet, so don't stress out on that part! (grin)