There are a couple of ways we scale better. On the pure data delivery side, we do it by breaking the file into, say, a thousand pieces, so delivery is not bottlenecking on the slowest agents. We allow the system to actually aggregate the bandwidth of hundreds or thousands of low-bandwidth agents. It's like a swarm of ants; we're trying to move data from point A to point B, and if you've got ants and gang up a million of them, you can move a lot of stuff. So in Mojo Nation, instead of trying to push a 4-gigabyte file through somebody's 128K[bps] uplink, you break it up into 4,000 or 4 million chunks, then have lots of agents -- on their DSL lines or 14.4K[bps] modems -- send you part of it.

The other thing we do to help it scale is try to use economic incentives. If at some future point it becomes economically viable for an agent to dedicate more bandwidth, then it will do so. If there is a market need and somebody can fill it, they will earn more Mojo and so they'll do it. So we basically let market forces decide what Mojo Nation needs at that point.

People on the Net have gotten used to getting their music and other files for free. Do you think people will really embrace the idea of payment and sign-up that is central to Mojo Nation?

The system is not designed to be as easy as others because it was honestly designed to scale up and work in the long term. But we know the market works. We've basically created a marketplace for information, a place where these agents can buy and sell on behalf of users or a company or whatever. What we're trying to say is that you will want to be assimilated eventually. Our system is resistant to parasites and it scales up well. Most of the other systems out there don't do that. We're just waiting for them to collapse. You're right, it is a little tricky. But for most users, it's just as easy as pointing their Web browser at the local agent and saying, "Go get it."

Does the person who originally uploads a file to the network get paid?

No.

Then how do content creators get paid -- what's the incentive for them?

You can earn revenue by publishing through a feature we haven't added yet, which is basically tipping. When you publish you can say, Here's a digital signature on this map: I published this file; this is who to give the tip to.

It lets users develop a reputation as a source of good information, an authority -- or perhaps a record label would publish a song with a signature from Universal Music saying, Yes, this is the official file we released.

What about the potential legal pitfalls? Napster, for example, is defending itself in court by arguing that its users engage in noncommercial sharing, but because of the way Mojo Nation works, this defense wouldn't seem to apply. What makes you think the labels won't come after you?

It's possible that they will come after us. But we spent a lot of money on very expensive lawyers over a year, in the early design phase, to make sure that we're covered by existing laws. The system is designed so that, for example, we fit under the safe harbor provisions of the Digital Millennium Copyright Act. And there are substantial non-infringing uses of Mojo Nation. It's used for online backups, for publication of unpopular Web sites. By being content-blind we can also be much more flexible than Napster, for example.

Napster, for example, has the problem of not being able to get rid of the files even if it wanted to. They couldn't play fair just because of the way their system was designed. With Mojo Nation, if a content owner comes to us and says, "Hey, someone has published my Britney Spears track and here's the blocks and the map," we'll say, "Well, you're correct, and according to the DMCA, we'll take it off of our servers." We'll remove those blocks and publish them as bad blocks. Everyone who subscribes to that list could say they won't traffic in that.

That wouldn't stop someone from putting up another Britney Spears song, though, right?

Right, it wouldn't. But that's not our problem. Remember, these content maps have to be disseminated to the users, so if the user submits that to a content tracker that searches specifically for MP3s, then that content tracker would have to follow the laws of his or her own jurisdiction. So if someone were running an MP3 content tracker -- which is one of the software agents that all users can run -- they would be in the same position as Napster is now. Functionally, that's what these content trackers do; they act as distributed index services and directories.

Then you're pushing off liability onto your users?

Exactly. We say that users are grown-ups and they can make their own choices.

When you say that creators of content might be compensated by tipping, aren't you making a kind of optimistic assumption about how people will behave? Users don't have to tip.

Well, part of the thing with peer-to-peer is that the user has complete control over the system. You can't force someone to do something. What you can do is make it easy and possible for them to do the right thing. Mojo Nation is designed assuming that what we have to do is build a system that works, so that people can effectively tip those who they think deserve compensation.

The Free Software Project
Read Andrew Leonard's book-in-progress on Linux and open source -- and post your comments.