IPv6 is for Porn?

April 16, 2007 Todd Underwood

I’ve written about IPv6 in the past—mostly to point out how little traction it has been getting and how unlikely it has become that IPv6 will be the next network layer protocol. A new project hopes to change all that. A hint of how they intend to accomplish this is available by noting that the same content can be found at http://www.ipv6porn.com/. The project describes it as follows:

We’re taking 10 gigabytes of the most popular “adult entertainment” videos from one of the largest subscription websites on the Internet, and giving away access to anyone who can connect to it via IPv6. No advertising, no subscriptions, no registration. If you access the site via IPv4, you get a primer on IPv6, instructions on how to set up IPv6 through your ISP, a list of ISPs that support IPv6 natively, and a discussion forum to share tips and troubleshooting. If you access the site via IPv6 you get instant access to “the goods”.

The founders of the project acknowledge one key IPv6 adoption barrier: the lack of content. There’s a chicken-and-egg cycle between content and users. Users want content to consume, but content wants users to consume it. As long as their is neither (in any numbers worth talking about) on any IPv6 network, there’s no reason for either users or content providers to migrate there. IPv6 is a network equivalent of the Bridge to Nowhere (another doozy from Alaska Senator Ted “The Internet is a series of tubes” Stevens). It’s expensive, goes nowhere and no one needs to use it. This project is an attempt to fix that.

But there are two big things missing from IPv6, and content is only one of them.

The other is any way to get from the IPv6 network back to the real Internet. Even if IPv6 networks have valuable, important content and this drives some amount of usage, every single one of those users will still need to get a way to access Internet services as they exist today. No one should seriously expect users to do without Google, Yahoo, Orbitz, CNN, BBC, Etrade, www.xkcd.com or Reddit. It’s just not reasonable!

There have been several, abortive plans for how to solve this problem (and simultaneously solve some of the real scaling problems with IPv4 routing). So far, most of these have come to naught. But a new proposal called “LISP” has been getting some serious discussion. The proposal’s main disadvantage is its name. Although it spells out “Locator/ID Separation Protocol”, a good guideline in almost every aspect of computer science is to avoid name reuse.

Still, LISP (the networking one, not the functional computer language that most of us love to hate) has a lot to recommend it: it’s practical, it’s implementable at the edge without global coordination, it can be phased in, and it comes in various, increasingly more ambitious levels of implementation. It also takes a stab at the long-term reduction of the bloat in the global routing table. In short, it tries to be everything that IPv6 always needed and many of the things that the IPv4 Internet is starting to need.

LISP was not designed to solve the v6 to v4 transition problem. But it is one technology that could serve that role. From the draft:


Note that while the detailed protocol specification and examples in this document assume IP version 4 (IPv4), there is nothing in the design that precludes use of the same techniques and mechanisms for IPv6. It should be possible for IPv4 packets to use IPv6 RLOCs and for IPv6 EIDs to be mapped to IPv4.

I’m not anywhere near smart enough to figure out whether LISP or something like it is actually a solution to some of the woes that face us, or whether it’s a reasonable transition plan from IPv4 to IPv6, either. It looks much more promising to me than shim6 ever did. But the LISP draft only came out just before the APRICOT meeting in Bali that I attended. Although there was some discussion of the overall problems associated with v4 table bloat in Bali, there was little or no detailed discussion of LISP or other proposed solutions to the ID/Locator split approach. And there was essentially no discussion of how something like LISP might be used as a transition vehicle from one network layer to another. It’s definitely worth watching, though.

Although porn for v6 is funny and intriguing, I suspect that this project, like those that came before it, is likely to fail. IPv6 networks still don’t make much rational, economic sense in the commercial sphere (with the only exception that I’ve found being something like the private IPv6 deployment at Comcast which is done for addressing of the infrastructure, rather than for endstation addressing for Internet access.)

If you find the naked people at http://www.ipv6experiment.com/ so compelling that you add v6 into production on your network (or even just your computer), please do let me know.

The post IPv6 is for Porn? appeared first on Dyn Research.

Read more...

Previous Article
Rough Day at Cogent Part 1: Cogent Depeers Someone?

These are the signs of the apocalypse: A worldwide earthquake, the sun...

Next Article
The Shape of Disaster on the Net
The Shape of Disaster on the Net

Network Outages Caused By Taiwan Earthquakes The quakes that damaged seven...