Tim O'Reilly's Post-Apocalyptic Pants2Pants

by Don Marti

Attendees at the conference included hackers from some of the cutting-edge crypto-elite trust-no-one projects such as Freenet, Free Haven and Mojo Nation (plus people from espra.net, who were searching for an Ethernet connection to finish their software for release during the conference); the inevitable venture capitalists and marketing people in search of business models; and reporters, some clueful, others just looking for whatever the Napster users are going to go use now that napster.com is Dead. The media line was as long as the regular attendee line at registration, for "Bob"'s sake.

Now, even though I'm writing this in the secret Linux Journal bunker in Seattle, I can still hear all you Old Un*x Farts out there scoffing in your mighty beards. Peer-to-peer is just end-to-end for people who don't remember end-to-end, right?

Not quite. Peer-to-peer deep thinker and bon-mot-meister Clay Shirky defined P2P with two criteria. First, a P2P application "aggregates resources at the network's edge". There wasn't a "network's edge" back in the day of end-to-end, when herds of wild IP addresses roamed the plains, and everything on the Net had a static IP address and was available all the time.

And, back in the day, everything had a True Name, thanks to DNS. The other part of Clay's P2P definition is that a P2P application does its own namespace, independent of DNS, because you can't just look up an ever-shifting Edge PC in DNS.

Edge PCs come in two flavors: home and office. The main difference is home PCs are more available to participate in peer-to-peer than office PCs cloaked behind Network Address Translation schemes and firewalls. At the opening panel, Ray Ozzie of Groove Networks said, "the Net is fundamentally broken right now because of NAT and firewalls." Corporate networks have plenty of bandwidth and plenty of unused edge resources, but most of their boxes aren't available to accept incoming TCP connections. At a panel called "Dealing with NATs and Firewalls", Ozzie added, "There are applications emerging that need the original design model of the Internet," and called for a new layer of standards to allow for transfer of "XML packets" between peers, even if both are behind NAT.

Firewalls have a security role, but in an increasingly complex business world, it's unrealistic to divide everyone into "us" who can see everything, and "them" who can't. Christian Huitema, of the Internet Society, said, "the idea that you can have a single drawbridge is pretty much fantasy", and "the function of the firewall is going to migrate into the operating system." (We pause now for a big "Well, Duh" from any Old Un*x Farts still reading. If anyone wants to bash the security models of proprietary OSes, please take it outside.) Even though firewalls may be bogus, any peer-to-peer system that wants to work on corporate networks is going to have to develop some trust among firewall-loving IS Department managers. Ray Ozzie and Groove certainly have a hard sell ahead of them.

Firewalls are just an annoyance to sneak around, and in the long run, no defensive line can hold against a determined invader. No big deal. But anyone trying to make a business out of peer-to-peer will have to face the two mighty Enemies of the Internet: software patents and Digital Rights Management.

If you thought the Web was a patent minefield, welcome to the rabid-weasel-infested nuclear minefield of P2P. At his talk on patents, Greg Aharonian said, "Anyone who figures out how to make money in P2P is going to get letters in the mail, saying `You owe us some money.'" Peer-to-peer patents go back more than 20 years and hundreds are in force now. Not that they're valid or anything -- the term "peer-to-peer" first shows up in patents in the late 1970s. "Pretty much all the fundamental technology has been published many times over", Aharonian said. However, the mere fact that the patents are bogus doesn't mean the patent holders won't put you out of business.

Even though Aharonian says that "the majority of software patents are invalid", and questions whether any of the steps in peer-to-peer software development have been large enough to merit a patent, he still hasn't come out against software patents in general, which is kind of like being in favor of allowing mutagenic chemicals in food because not all mutations are harmful.

The final layer of concrete in the containment structure over the vitrified capsule over the last nail in the coffin of the P2P business is, of course, what Professor Lawrence Lessig called the "weird exception to the First Amendment", the Digital Millennium Copyright Act, which sanctifies legal attacks on any technology of which large copyright-holding corporations do not approve, and enshrines any Digital Rights Management technology, even when it prevents legitimate use of DRM-infected content. "It's not the Framers' Constitution", Lessig said. "It's not even your father's Constitution." Courts, Lessig said, are upholding a new, dangerous legal right--any copyright-holding corporation can force developers of any new technology to architect their project to prevent copyright infringement from the start, whether or not they are left with a working system.

John Perry Barlow's answer to that, in a panel discussion following Lessig's talk, was "The only way to deal with law in cyberspace is to ignore it. Go out and develop whatever you please, ignoring what the law says." Dan Gillmor from the San Jose Mercury News tried to bring some moderation to the discussion, asking "Why is the technology industry not fighting this?" but the self-described "addled old hippie", Barlow, answered, "Congress has been bought and paid for by the content industry." But corruption isn't the whole problem. The public doesn't see hackers as a lovable Bo and Luke to Jack Valenti's Boss Hogg. EFF and other freedom-lovers have "failed in explaining the issues to the general public", Lessig said.

Clay Shirky and Tim O'Reilly joined Gillmor in sweet moderate happytalk, but the facts on the ground are that DMCA thugs are kicking hackers' doors down, not the other way around. Gillmor said if he had to choose between the end of fair use and libraries, and the end of his own copyrighted-newspaper business model, he would take the latter and find a new job, which is damn cool but doesn't change things much.

The fact that there's no gold to be had in P2P didn't stop the productive technical discussion among the hackers. After all, being driven underground and living like rabbits in Watership Down, network-wise, is just a way to test how well your crypto infrastructure and distributed reputation metric system protects you from the elil. In the hallways, memes, solutions, math and hacks migrated from project to project like plasmids--as soon as people figured out each other's terminology. One hacker, Brandon Wiley of Freenet, explained, both at a talk and to people in the halls, how and why to put all kinds of regular Internet applications on top of Freenet. One popular Freenet game could get a critical mass of nodes running, and after that, every application will be able to use the same infrastructure--whether it's for software distribution, chess, news or whatever.

So, hackers, what did we learn in San Francisco? First, stay poor so the patent weasels won't have the incentive to sue you. Second, get or stay pseudonymous, because the movie weasels will stop at nothing to try to shut you down, money or no money, non-infringing uses or no non-infringing uses. Third, stay open source, so when any VC weasels dumb enough to throw money at you wise up and bail, you can jump in the escape pod with your source code and try to find a habitable planet with oxygen and a day job. Oh, and if you go to a conference, tip the bartender well, so you keep getting Anchor Steam after everyone else is reduced to Budweiser. So Tim, nice pants, I mean good conference. The lessons may have been a little post-apocalyptic for most of your audience, but hey, see you at the next one.

Load Disqus comments