IPv6: What Are They Really Thinking?
IPv6 is here. Well, it has been for 16 years now, but aptly enough, World IPv6 Day is all about its coming of age. No candles or shoe ceremonies at this event, simply a lot of big names with seemingly altruistic and forward-looking intent. Want to see if your setup can handle it? Probably not, unless you just bought a cutting-edge router, and connect to an ISP with widespread IPv6 deployment (there aren't any, BTW, but Comcast, Time Warner, Sprint, and AT&T are trying. No word from Cox, Cablevision, Verizon, or T-Mobile).
Still, IPv6 is an inevitable reality, although it will coexist for the time being with IPv4. As more and more IPv6 sites and hardware become the norm, IPv4 will become increasingly isolated and detrimental to stubborn businesses. Many companies are convinced that the time is now, and are helping The Internet Society push for IPv6 deployment and awareness.
Okay, but why? Well, the biggest reason is simply that we're running out of IPv4 addresses. That alone is worth transitioning to a more capacious specification. Get used to the word "undecillion," as in "IPv6 allows for 340 undecillion addresses, instead of a measly 4 billion or so with IPv4." The fundamental benefit is that all of those devices that are currently behind a NAT router will now be able to connect directly. This frees up all of the resources wasted in directing local network traffic through the bottleneck of a shared IP address.
In general, this is great for new and newly expanding businesses, who will benefit from lower costs and more efficient connections. The advantage for existing business depends on how extensive their deployed networks may be; anyone who faces a costly IPv6 upgrade to supplement their inevitably obsolete IPv4 websites and hardware will resent the bite it takes out of manpower and revenue.
Latency has always been a thorn in the side of specific real-time online activities, especially gaming. IPv6's direct one-to-one connection means a drastic reduction in latency, one of the reasons Steam and Valve (Counter Strike) are behind the IPv6 push. Oddly enough, Blizzard (World of Warcraft, Starcraft), is silent on the issue, as is EA Games (Call of Duty). Also unspoken are both OnLive and Nvidia, whose revolutionary Virtual Network gaming ventures rely even more on reduced latency.
However, gaming is not the only game in town. The prospect of one-to-one connections make YouTube and Netflix eager about the potential for IPv6 to allow for more efficient video streaming. The publicity is also not bad for traffic; sites of companies mentioned in the recent IPv6 coverage have seen a spike in visitors. On the other hand, the traffic may have been mainly comprised of nervous customers hoping that the sites would still work.
Even services which depend less on real-time latency stand to gain a huge boost in reducing the time and complexity of device-to-device communication. Owen DeLong of Hurricane Electric used TiVo as a prime example, saying that remote scheduling now could take up to 2 hours to complete, whereas with an app-driven IPv6 connection, it could take a mere fraction of that. Sounds like an obvious FTW situation for anyone who is in the business of selling Internet-connected devices, doesn't it? Funny you should say that...
We do have a pretty good idea of who might not want IPv6 to come around too quickly. The list includes Apple, IBM, Haliburton, Xerox, pharma giants Merck and Eli Lily -- all of whom own sizable amounts of unused IPv4 addresses that get more valuable as supply gets more scarce. Apple is perhaps the single most surprising of these, considering that the company's business depends on exactly the kind of always-connected, always-trackable, fully integrated devices that IPv6 makes more possible than ever. WWSJD?
By the way, AT&T is on the list of IPv4 address hoarders as well. However, with a growing base of smartphone and tablet users, they also stand to benefit from the IPv6 rollout. We can assume that AT&T executives fell into opposing camps over this issue.
Okay, so there isn't much question about network hardware manufacturers; they're as giddy as TV manufacturers and cable providers when everybody found out that their analog televisions were about to become worthless. Cisco (including Linksys), D-Link, Netgear and others can now sell IPv6 hardware to existing customers who would have been happy with their IPv4 equipment for many years to come. Huawei in particular offers numerous IPv6 solutions, including a recent deal with China Telecom and SingTel group on the provider side. However, the transition is going to be slow, which benefits those who want to stick with IPv4 for a while. So obviously it's not all about an enforced upgrade.
Google is heavily involved, and not simply because Apple isn't. In fact, both iOS and Android devices tend to support IPv6 natively, and the spectacular growth of the mobile consumer base is a major impetus for the addition of many new addresses. This presents its own problems, because most of the Internet is still made up of IPv4 servers. But Google has the luxury of implementing both simultaneously, so they're covered whether the transition is fast or slow.
Facebook's stake is less obvious. With a shaky IPO and revenue pressure from increasingly involved shareholders, you might expect that the last thing Facebook would want is an investment in an expensive project with no immediate monetization potential. On the other hand, Facebook is all about personal information, and the more elements of a person's life that connect directly to Facebook's servers, the more efficient Facebook will be about delivering useful data to advertising partners. This is especially true considering the increasing broadband penetration of developing parts of the world -- those self same parts which provide Facebook with its most opportunity for growth.
Which brings us to privacy and security concerns. You'll hear IPv6 evangelists touting the baked-in encryption that IPv6 offers, but they'll say less about some inconvenient truths -- such as the fact that Internet Protocol Security (IPsec) was such a good idea that IPsec has been bolted on to IPv4 for some time. And it should go without saying that all of that now-unnecessary NAT hardware tends to provide firewalls and other security measures that will no longer be a protective gateway. So where are the crucial security enterprise IPv6 sponsors?
On top of that, the direct addressing means that physical location just got a lot more precise. When I connect to my ISP, my IP address can only be traced to the nearest node -- which means that a snoop might be able to tell what general part of town I'm in, but not the exact place that I'm connecting. IPv6 is a boon for anyone who can benefit from knowing precisely where any given device might be, That's great if you want to find a stolen smartphone -- but not so great if you're the author of a blog that has been criticizing your country's totalitarian dictator.
The hardest part of coming to grips with IPv6 is the inescapable irony: the better it works, the less you should notice it at all. Plenty of emerging ventures will stand to gain from access to a vast new supply of cheap IP addresses, and everyone gets a potential performance boost and integrated security. It seems like a clear win across the board -- unless you're a company that faces rapid obsolescence of your network, or of your valuable IPv4 address holdings.
Without knowing the "backroom" conversations had by current IPv6 proponents, a bit of skeptical paranoia may be forgiven. We've all seen examples of urgent needs and widely advantageous advances that have been placed on the back burner due to an unfavorable cost / benefit analysis.
I hate leaving the reader with more questions than answers, but perhaps someone can provide the link that I'm missing. Why do you suppose IPv6 is receiving this level of support and attention -- and why are so many silently waiting by the sidelines, when they seemingly stand to benefit more than the supporters?