IPv6: What’s the Rush?

Q: I saw a recent article about IP version 6 and how everybody needed to start changing their addresses over in order for the Internet to continue working. My boss started asking me questions about this, and I don’t know what kind of answer to give.

None of my friends seem to be concerned or interested in changing, though, so is it really a problem?#

Answer:

If you look hard enough, the sky really is falling. Or perhaps the Earth is rising. I haven’t figured out which one yet.

The bottom line is, “it depends” — which is a standard Cisco answer for things, but it seems to apply to everyone in the case of IPv6. So let’s look at a few things.

What problems are we having with IPv4 that may require us to change? Namely, there are not enough addresses. Part of this problem, of course, was that in the beginning, people were handing out IP addresses like they were candy and in class-full boundaries. So absolutely there weren’t enough!

I have clients who have /16 address space that they honestly use perhaps 50 to100 devices on. But there isn’t any incentive for them to give things back, so they aren’t giving them up. And regional Internet registries (RIRs), like the American Registry for Internet Numbers (ARIN), have made things incredibly complicated and convoluted for transferring these older portable addresses, so selling them is hardly worthwhile. Consequently, people just hang on to them for no good reason while others struggle.

Even with the advent of Variable Length Subnet Masking (VLSM) and the ability to assign addresses in smaller allocations, we still saw the writing on the wall: The number of people and businesses on the Internet will increase all the time. That was just a way to delay the inevitable.

IPv6 offers us significantly more addresses than before. We would move from a 32-bit addressing scheme (4.2 billion varieties) to a 128-bit addressing scheme (lots of varieties!). IPv6 promises us somewhere in the neighborhood of 340 dodecillion usable addresses, in case you really cared there. Google that one!

But we have other things that are staving off a mass exodus, the first one being humankind’s resistance to change. Oh, yeah, and the money thing. We won’t spend money unless we have to.

The other big thing is Network Address Translation (NAT). There’s actually a large debate these days about the viability and proximity of IPv6 usage because more and more people are using private addressing (RFC1918), and the quality of NAT devices and firewalls is much better today than it ever has been before.

Newer items also cropping up are Application Layer Gateways (ALGs), Layer 7 Firewalls or super proxy servers. Each of these things, though, does not obviate the need for more addresses; it just delays our need to change.

After we get past the desire, motivation and money allocation, we get into the technical aspects. Is all of your networking equipment capable of IPv6 configuration? How about your security equipment? How about your operating systems and applications? How much of a deployment you will want to explore will depend on many of these. There are many ways to do NAT between IPv4 and IPv6, so that your internal systems may not need to migrate (or at least not as quickly) and your outside-facing configuration changes over.

How about your IT and networking staff? Has everyone been trained on IPv6? There’s a lot more involved than just extra bits. There are many nuances and differences in protocol and particulars along the way. DNS changes, DHCP changes…the world is a different place!

All of those things are important to think about. Now may be a good time to start evaluating your readiness. But in the end, as a normal enterprise (small to medium business), you likely aren’t going to change any faster than what your upstream provider is doing or requires. So what I would do is ask them.

A couple of smaller ISPs I’ve talked to aren’t even in the planning phases for IPv6 yet. Many are fairly well along in testing, and some even have implementations underway. IPv6 does not need to eliminate IPv4 initially, so it makes things nicer for planning an implementation. But it’s still something that takes a lot of detailed planning before jumping into it.

Jeff Doyle wrote a blog entry not long ago about this where he pointed out a similar idea: Much of the deployment depends on what people are ready for. There is no magic to making it happen. Eventually, yes, we will run out of IPv4 addresses, but the question is when.

ARIN (and other RIRs) are recommending people to push for IPv6 going forward. Most will stop handing out IPv4 allocations in 2009 (get ‘em while you can!). But who knows what trends in reallocation or security and NAT will change in the next couple of years, which may push things off further.

The bottom line is that, most likely, there isn’t any dire or immediate need to change everything out. On the other hand, if your upper management is interested and/or concerned, now may be a great time for any budget approval you need to change things! But you should at least consider your five-year planning phase. How many addresses will you need for your entire company network? Can you get those now? Do you need to get those now?

In planning for IPv6, don’t forget the human part. After all of the pieces (network, applications, OS, etc.) are done, do you have enough people with enough knowledge to manage and design things? Now may be a good time for some training!

There are various vendor-based training courses out there on IPv6. There are also different presentations at meetings like the North America Network Operators Group (NANOG) regarding important topics like this. I would advise you to take the time to research things with regard to your organization and what it would take to migrate over. No rush, but having the time to plan is good!Hope that helps :)

Press our buttons!
  • Digg
  • Sphinn
  • del.icio.us
  • Facebook
  • Mixx
  • Google
  • Furl
  • Live
  • Reddit
  • StumbleUpon
  • Technorati
  • TwitThis
Related posts: