Should an IPv6 deployment mirror an IPv4 strategy?
Some people assume that an IPv6 deployment is just like that of an IPv4. This can be a good thing because it doesn't scare people away, but, at the same time, it steers companies into using a matching strategy when deploying IPv6 -- meaning they try to replicate what they did in IPv4. It's a terrible idea to attempt to match the IPv4 address space into IPv6. Don't assume that features, designs and architectures all match up because you will miss the opportunity to do things the right way, and you will send the wrong message to vendors as well.
Here's what I recommend. If you know what you want from your IT environment, do not focus exclusively on parity. I would rather have folks go and ask vendors to build new capabilities than just push for parity for the sake of parity.
Second, it's important to develop a deep understanding of IPv6. After you familiarize yourself with the first wave of IPv6, I suggest taking a step back and exploring in more detail the technology's control plane, functionality and capabilities. The second time around is when you will understand a lot more about the fundamental assumptions that are made within the protocol, along with how the protocol is implemented. Once this process is complete, tie in IPv6 and your understanding of IPv6 with other technologies that are relevant and key in your IT environment -- the cloud is one example.
You will subsequently be well-equipped to leverage your IPv6 knowledge into the cloud. You'll be able to come up with a design for the cloud and its infrastructure that will actually scale much better than IPv4 allows. You will also have the ability to drive the right type of requirements -- including IP Address Management (IPAM) requirements that not only reflect IPv6 needs but also tie in cloud needs.
Editor's note: With permission of the presenter, this expert answer was excerpted from a recorded presentation given by Popoviciu in 2011 at gogoNET LIVE!, an IPv6-centric conference sponsored by gogo6. View the full presentation here.
This was first published in May 2013