In Part 1 we discussed how to turn off ISATAP on Windows host—which is a great idea. Turning off unnecessary components of your network simplifies everything. But ISATAP can be useful in certain scenarios. For instance, if you want to test an application on IPv6 you clearly don’t want to turn on IPv6 everywhere and just cross your fingers. With ISATAP you could just enable IPv6 on a small group of beta testers. With the exception of one Cisco router (discussed below) you wouldn’t turn on IPv6 anywhere else on your network.
Yes, you should run your network on network gear. I know, crazy, right!?! So, I wouldn’t recommend using a Windows Server as an ISATAP router, which is what you’ll get if you start trying to google “isatap windows setup.” For all those Windows admins out there, we’re going to use a piece of purpose-built network equipment from Cisco, a popular vendor out there in the networking field (the stuff on the other side of the Ethernet cord you plug in.)
In our setup we have server at 220.127.116.11 and two clients that reach the server over an arbitrarily large IPv4 routed network.
Our Cisco router config will have the following.
interface Loopback1 description < ISATAP Network 1 > ip address 203.0.113.30 255.255.255.255 interface Tunnel1 description < ISATAP Network 1 > no ip address no ip redirects ipv6 enable tunnel source Loopback1 tunnel mode ipv6ip isatap
I use a loopback to source my ISATAP multi-point tunnel interface. This makes it easy to multi-home the router. The ISATAP doesn’t rely on either physical interface to be up; as long as the loopback is reachable (I redistribute it into OSPF so the network knows how to reach it) the ISATAP overlay will be up and running.
Yes, I know there is more config that we will need to make a truly usable system. I’m building suspense!
I’ve been lagging behind in my blogging (Part 1 was posted in April!) so I’m going to try and keep things shorter. In Part 3 we will configure the Windows hosts to connect to the ISATAP overlay and form a workable IPv6 network.