Friday, January 13, 2012

The IPv6 Message

I've been working with IPv6 for a long time - I started when all we had were 6bone tunnels, and to use them you had to get EFT code for your Cisco, compile your own Linux kernel and find v6-enabling patches for your applications. I've never contributed to the core IPv6 development effort, but for a significant part of that time, I've taken on the role of IPv6 evangelist within my networking community. I've taught workshops, done demonstrations, written papers and blog articles and made presentations in front of hundreds of people. Many of them were skeptical about what I was telling them, concerned about the implications of acting as much as of waiting. In all that time, though, I've never had such a convergence of events as happened in the last month.

On three separate occasions since early December I've heard people say something that I never expected - smart people, who are involved in networking and information technology, who make decisions for their organizations and are, or ought to be, in touch with what's going on in the larger networking world. I'm paraphrasing slightly, though I think you'll get the idea:

"I don't think you'll have to worry about IPv6 any time soon. Maybe in ten years."


"I keep hearing a lot of hype about IPv6. I think it's just another fad, like ATM."


"My company isn't committing to IPv6 yet, since we don't know if that will be the direction the Internet takes."

To say that I was taken aback would be an understatement. I didn't challenge the speakers - there was nothing to be gained by debating any of them. I haven't forgotten what they said, though, and it bothers me.

I don't think that I'm an IPv6 zealot, but I can't see any future for the Internet that doesn't involve v6. We can debate whether the transition should have happened quicker, whether a particular design decision was optimal. I have had many conversations about the right way to transition and the appropriate timing. Never, though, has there been any question about its inevitability.

If there is an alternative to IPv6 I think that I would have heard about it; if I've missed a memo I do hope that someone will enlighten me. I really don't think that will happen. I'm concerned, therefore, that intelligent, involved, technical people would hold these opinions. In fact, two months ago you couldn't have convinced me that anyone did; that's a big part of what makes me worried about this.

So that leads me to ask some questions. What are we, IPv6 enthusiasts and evangelists, doing wrong? What have we missed? Have we been so close to this process for so long that we've forgotten to talk about things we now take for granted, including the fact (and I do think it's a fact) that IPv6 is absolutely inevitable? Did we cry wolf too many times during the last decade, warning of seemingly imminent problems that didn't arrive on schedule, until those warnings lost credibility?

And how can we fix it, educate those who have this misapprehension, without alienating them or making them feel like fools? I think we owe it to them, because we're the ones who have been carrying the message all along. And if we wait for the future to show them the error of their ways it'll only mean more work for everyone.

2 comments:

  1. "What are we, IPv6 enthusiasts and evangelists, doing wrong?"

    I get the same feeling in my stomach for DNSSEC, and you could probably take this posting and replace "IPv6" by "DNSSEC". (*sighs*).

    ReplyDelete
    Replies
    1. I haven't been evangelizing DNSSEC for nearly as long as IPv6 - mostly because it wasn't ready - but nowadays I try to push both of them whenever I can. It's an interesting contrast: on the one hand DNSSEC doesn't have the inevitability of IPv6, but it offers tangible benefits with the prospect for more on the horizon (DANE, for example). DNSSEC is also much better bounded, in terms of its impact on the organization. It is tough to find benefits for IPv6 adoption if you already have enough IPv4 space, and the impact is huge. And yet v6 is the one that we *must* do.

      Delete