Consider the fundamental ideas behind P2P.
In most Internet applications, a client contacts a server for some content (e.g. a web server), or a service (e.g. an online bank) or to communicate (e.g. a videoconference). The beauty of the Internet is it is built around the end-to-end principle: everybody can contact everybody else, in a uniform fashion. Everybody can be a client, everybody can be a server, and everybody can be both, at the same time. They can all be peers.
P2P technology comes in a variety of forms.
The seti@home project, for example, consists of a server which create parcels of work, that downloads work to clients such as your desktop PC, and returns the results to the server. As there is no communication between the peers seti@home isn't really P2P in its purest form, but it's a start to sharing resources such as storage and computing power.
Napster consists of a server providing effectively a directory service. Your client contacts that server, tells it what you offer to the rest of the Internet, and then lets you query it for what others are offering to you. If you request a resource you are placed in contact with the peers who have what you are after, and communicate with them directly.
It is possible to achieve the same thing by installing a simple web server package on everybody's desktop computer, and registering your content on somebody's central web server.
Napster isn't really any smarter than that - it just did it in a more user-friendly way.
You can take this model further, by distributing the directory servers, or by making them part of the network of peers itself. This is the approach of e.g. Gnutella, amongst others. The advantage is that the system becomes more robust against failures on networks or the servers themselves. Ultimately, everybody is an equal peer.
Can this technology be used for something besides sharing music, movies and software?
Absolutely.
Even in a very rudimentary fashion, sharing content in an efficient manner is vital to a range of new computing challenges.
Projects like Seti@home, and more generally grid and Web services all require a standard interface and provide access to a massive networked infrastructure. One example is distributed computing, which harnesses the power of a large number of network-connected computers, allowing the user to identify the best computers to run their problem on, based on cost, performance, location, availability, or a range of other issues, and it can even change these decisions during the computation all thanks to P2P computing technology.
Data grids and P2P content sharing offer efficient access to data distributed around the network, which is needed by distributed computing activities. Consider a simulation merging weather data from a central bureau, with local geographic data from various councils, and remote sensing information from satellites, to provide rapid predictions on the behaviour of actual bushfires.
P2P visualisation and collaboration systems enable users to visualise results and discuss them with their colleagues in real time.
One day, even a simple activity like booking your next trip through a travel agent will ride on top of a set of web or grid services, with underlying P2P technologies.
Research is also being carried out aimed at using P2P-based content sharing, to prevent spam (unsolicited commercial email), others are using it as a way to share documents or other content that might be politically sensitive in some countries. Another group of projects are looking at reducing the huge amount of replicated content across hard disks within an organisation (no matter how cheap a huge disk is, it is always cheaper and easier to overflow it).
P2P computing offers more robust computing environments, on a much larger scale, and at greatly improved price/performance ratios, compared to monolithic systems, for many applications.
Given its perceived misuse today, can P2P content-sharing be stopped?
While I can certainly see its misuse, I can also see various problems with trying to stop it.
The first one I see is a clear definition. You can only effectively legislate against something that you can clearly define.
A set of machines running web servers with linked pages, a search engine like google, a mesh of email servers who can contact each other over the Internet, a series of IRC servers, web caches and proxies, these are all P2P content-sharing systems. And it would not take a lot of effort to build a Napster-like system on any of these. In fact, things like Napster have been round since long before Napster itself came along.
Legislating against "just" P2P would probably block many other, legitimate, uses of the technology. At the same time trying to word around these legitimate uses would probably create too many loopholes.
Next, such legislation would be basically unenforceable. Napster was easy to shutdown because the technology was too simple - take out the central server, you take down Napster. Projects like Publius, on the other hand, set out to have no central ownership, no central resource, users do not know what content is stored on their machine and no complete item of content is stored there unless you yourself downloaded it. It's even encrypted end-to-end, and anonymised - so its impossible to tell where content is coming from, or going to.
It is possible to attack the P2P content-sharing technology itself. Some industry groups have suggested flooding P2P communities with fake content, or even malevolent content (e.g. viruses). Some people have suggested this is already happening. I don't believe this approach will work for very long. There are many people working on ways to improve the protocols and the software, and it will be very easy to detect, and block, attempts such as these. We need this to protect legitimate uses of P2P technologies such as grid computing and web services.
There are a large number of issues around rights-management of assets that can be stored and transferred digitally, which need to be solved. I think P2P content sharing shouldn't be seen as a (or even "the") problem, but simply a technology that has highlighted these issues. It's not the only one. I believe Napster's success reflected the power of mp3 compression and home/school/office broadband networking as much as user-friendly p2p content sharing.
Should we try and stop those as well?
Markus Buchhorn works at the ANU. Opinions expressed in this article are his own, and do not necessarily reflect the opinions of his employers, or anybody else for that matter.
