I was doing some reading here, and it got me to thinking. IPv6 has been around for a while, so why isn’t IPv6 widely deployed? Why don’t I see blog articles from the trenches about experiences learned during IPv6 deployments? Why aren’t carriers beating down my door to sell me IPv6 connectivity (goodness knows they have my number)? The reason is simple: the vast majority of us have no need to deploy IPv6. Therefore, we aren’t. Sure, we are out of public IPv4 address space. We’ve heard all about it. But just because “we” are, doesn’t mean that “I” am…and probably not you either.
Moving to IPv6 isn’t as simple as checking a box to support another octet’s worth of address space (which could have taken us rather a long way, if you think about it). IPv6 is effectively a whole new way to communicate, and so there’s a learning curve for busy geeks. There’s a dual-stack requirement, because IPv6 is not backwards compatible with IPv4. Not all hardware and software offers the same level of IPv6 functionality. Networking hardware has had a hard time with scale because of the sheer immensity of IPv6 address space – an issue more for the carriers than the rest of us little people, but a point nonetheless. IPv6 is a pain in our collective fanny, and frankly, we have more immediate problems to resolve. Like vMotion between data centers over LFNs, that kind of thing. Stuff that if we can work out, helps our businesses. IPv6 doesn’t really fix anything for us…and we instinctively know that once we get rolling with IPv6, new problems are bound to show up. Probably annoying, buggy, complicated ones involving long vendor phone calls, IVR queues, and help desks with little help to offer, as we’ll all be in this IPv6 deployment thing together.
The transition to IPv4 was driven by a strong impetus: the Internet had arrived. Get connected, or get left behind. Businesses quickly saw the potential value in web sites and e-mail, and so the quest for money drove the Internet’s growth. Microsoft, Linux, Novell, and Apple made it easy for us to transition our internal networks away from dual-stack by allowing us to connect to their services natively over IPv4. And don’t tell me Apple and Novell didn’t choke on killing AppleTalk and IPX – but they did what had to be done in the face of the tsunami. (I imagine Microsoft had less of an issue walking away from NetBEUI – a bit of an embarrassment, really.) The geeks sorted it out, consultants made some money, RFC1918 plus NAT became second nature, we figured out that open SMTP relays were bad, and at the end of the day, we’d all sorted out IPv4.
What’s the kick in my pants to deploy IPv6 at my enterprise? Does my company need IPv6 to do business with other companies? No. Are there mission-critical resources available to my business only through IPv6? Nope. Will my company reach a new market upon deploying IPv6? No again. The only drive I really have is that I’m a geek, and it seems like IPv6 is probably coming. Just maybe I should get out in front of IPv6 as a leader instead of letting IPv6 run me over like a drooling n00b. But I have to say – I’m not so sure it’s getting here anytime soon. And I really have to sort out this whole virtualization thing, because vendors keep shoving network functionality into that darned hypervisor…oh wait, BRB, the ticketing system just sent me a love letter.
Update later in the day…this showed up on Google+, and I had to add it. Too funny. “NATs are good…”