My pinko commie P2P notions
Shawn Fanning wrote Napster, the world's first peer-to-peer application, so that students in his dorm could share music with each other. Ever since then, P2P file sharing has remained the killer application for digital music. First, to be fair, Napster is not a true P2P network. It uses a centralized server, which acts as a sort of air-traffic controller for the network. A true P2P network functions just like the brain. Its network of nodes doesn't need a centralized server to help make connections between the nodes. Instead, the nodes connect to each other autonomously, based on a set of rules that encourage the most efficient possible dissemination of data within the network. Just because this architecture has made inroads in the area of music sharing doesn't mean that it can't be used for all kinds of other things. I've come up with a few wild and crazy ideas of my own for using peer-to-peer technology to replace everything from our electricity grid to the bank system. While some of these schemes are more plausible than others, they all demonstrate how powerful the concept of P2P really is. Shockingly, all of these ideas are (gasp) very, very communistic in nature, which might say something about where technology's taking us.

P2P cell phones
When you have a conversation via a digital cell phone, your voice gets encoded into ones and zeros and sent to a nearby tower, which routes your voice through your wireless provider's network, to another tower, and finally to the cell phone of the person you're calling (where the ones and zeros get turned back into sound). It's actually a fairly simple process. But when we apply the peer-to-peer model to digital cell phones, things get interesting fast. Since many new digital phones have been incorporating GPS into their hardware, they could route those ones and zeros very efficiently from point A to point B. Basically, bits of your voice would be sent through a series of other cell phones until the data arrived at its destination. If the ones and zeros representing your voice were split into a number of different streams, sent via different paths, then reassembled on the phone of the person you were talking to, the data would get there even more efficiently, and security concerns would go out the window. Even if someone tried to intercept the conversations being transmitted through their phone, they'd be able to hear only a bunch of unintelligible syllables coming from many different people's conversations. Obviously, a scenario such as this would require some pretty elegant routing rules, and as a mere editor, I can't really comment on the technical plausibility of such a system. But while P2P seems to have been with us for ages due to the distortion of Internet time, it's actually still in its infancy, and I'd be willing to bet that one of the thousands of very talented minds working in the peer-to-peer area will eventually figure it out.

P2P Internet access
When I stepped outside to take a break from March's Digital Hollywood Broadband conference in San Jose, I ran into Cory Doctorow of OpenCola, who's working on a pretty cool peer-to-peer system himself. (OpenCola will be a network of folders that attract similar content from other folders.) We started talking about the potential of 802.11 wireless LAN base stations. Doctorow said that while driving around with an 802.11-enabled laptop in his car, he found that he could pick up Internet access from time to time when passing close enough to someone's 802.11 base station. We realized that if enough people installed one of these, the Internet could grow ad infinitum, without the help of any ISPs, phone companies, or any other service providers. Cory came up with an even wilder idea: installing an 802.11 base station in every car, then running a comprehensive anticollision program on the mobile network. You have to love it--using P2P to prevent P from running into P.

P2P processing and storage
Here at CNET, hundreds of people use hundreds of computers for hundreds of different reasons. But when it comes right down to it, they're all basically just using their computer processing power for performing tasks and disk space for storing data. I use a 733MHz HP Vectra with 28GB of disk space, but most of the time I don't use anywhere near its 750,000 cycles per second. And even with all of the MP3s that I've (ahem) legally purchased, I'm still using only 14.78GB of disk space (just more than half of what's available). I suspect that most of the other computers on our network also use only a fraction of their processing power and storage space. Since they're all hooked together on our Ethernet network, there could be a way for us to use much cheaper machines--say, something with 166MHz of processing power and 2GB of storage. Running a P2P computing network would let all the machines pool their processing power and storage space and share it all as a common resource. Even on those slow machines with minimal storage, each individual would feel as if they were working on a fast machine with plenty of disk space. Even though the network would have fewer cycles and less space in total, applying the P2P concept to the situation results in such efficiency that there'd be enough loaves and fishes for everyone.

P2P banking
It's fairly common knowledge that the money in our bank accounts isn't sitting in some vault somewhere. Banks make money by loaning your money out to other businesses and individuals at a higher interest rate than they give you for loaning your savings to them. Instead of using a centralized lending service to transfer money from investors to borrowers, a P2P network in this capacity would directly match up chunks of money with chunks of debt, minus the bank part of the equation. Borrowers would run a node specifying how much money they need, for how long, and at what interest rate. Similarly, investors would tell their software how much they want to lend, for how long, and at what rate. Even though the likelihood of a one-to-one match between an investor and a borrower would be very small, the system would be able to work out complex arrangements to make the whole thing add up. Borrowers might be loaning from 50 different people at 50 different rates that work out to be exactly the deal they're looking for when all the rates are averaged. Likewise, investors might be lending their money to 500 different borrowers at 500 different rates that average out to the figure they specified. The P2P architecture could supplant the banking industry as we know it, with money flowing more efficiently between borrowers and lenders. Banks would no longer be able to charge tolls on money as it passed through their system.

P2P auctions
These would work similarly to the banks, but since a one-to-one match between buyer and seller happens with every transaction, the system could be constructed in a much simpler way. Sellers would just have to enter their inventory, along with descriptions and photos, into a standardized seller's client for the P2P network. Buyers would be able to search for antiques the way they look for MP3s on Napster. When a match was made, the buyer and seller would work out payment, the same way they do on eBay or CNET Auctions. In fact, the service would feel very much like the auction sites of today, except the sellers wouldn't have to pay fees to list their inventories.

P2P electricity
If you install a wind turbine or solar-powered unit, you'll find that you can't produce a steady supply of power, because some days are sunny and windy while others are cloudy and still. Applying the peer-to-peer model to this problem would mean linking everyone's backyard power generators and installing meters at each node so that people received as much electricity as they produced (more or less). People who lived in locales with cloudy, still days would be borrowing from people who were having sunny, windy weather; when the elements reverse, so would the power supply. Essentially, the P2P aspect of this system would result in people being able to get a steady supply of power based on the average amount they produced. Taking it a step further, people in dark, still places would be able to pay a sort of licensing fee to a central pool, which would then be divided in a prorated fashion among nodes that were producing excess power.

Am I some kind of commie or what?
The peer-to-peer ideas that I've outlined above have one thing in common: they're all essentially communistic ideas. I don't mean that they're communist in any political way, since technology has no real political agenda of its own. (In fact, I'd argue that like living species, its only tangible goal is to further itself.) Rather, these P2P ideas are communist in the sense that they are all about individuals embracing, trusting, and helping each other based on their membership in a group. In fact, the only stumbling block between us and a P2P world could very well be the fact that the omnipotent, multinational corporations that could mobilize efforts of this scale might not want to see them happen (RIAA, anyone?). Perhaps that's why the open-source and P2P development communities already seem so intertwined. It's because only the P's in P2P have a significant interest in building these networks.

E-mail me your own wild and crazy P2P ideas.

Back to intro