Peer to peer technologies have a pretty bad name. The immediate association is with BitTorrent and pirated movies and ISP throttling but step back from that and you’ll discover that there are some interesting products turning up that take the old idea of the network being the computer and putting it to good use.
Probably one of the first peer to peer applications to get traction was SETI@Home which parceled out data from the Arecibo radio telescope for users to analyze in the hope of finding repeated patterns. They may still be searching for intelligent life in the universe, but the idea spread and a number of similar @Home projects developed their own architectures but all on the underlying premise – by getting users to donate CPU cycles they could contribute to the project – be it searching for aliens, a cure for cancer or the largest prime number.
Over time a number of these projects realized that having different runtimes and communications infrastructures was inefficient and didn’t help optimize the network effects of a peer to peer community and eventually BOINC evolved as an open source grid computing platform that in turn supported SETI@Home, Folding@Home, ClimatePrediction.Net and many others.
But searching for aliens isn’t the end of the platforms taking advantage of the power of peer to peer processing. There are two other platforms that have recently launched that are challenging the established thinking.
Search itself has traditionally been the remit of the companies with the big pockets. Microsoft, Google, Yahoo! and others dedicate millions of dollars to building and running the server farms that crawl the web, build the indexes and serve up the queries to their visitors and unless you’re pretty sure of your business model that’s going to be a scary market to break into. Unless your business model doesn’t need you to deploy all those servers or storage because your uses will.
Faroo, a German start-up, have exactly that model. Users run a small, lightweight application on their machine which serves three functions. The first is that it runs using idle CPU and network bandwidth to crawl the web building up an index and distributing it around other connected peers for optimal search performance. The second function is that responds to search queries – both yours and, if suitably configured, other users to give results from your local cache, the distributed index and various 3rd party search solutions. The third function is that it monitors where users are going on the web and uses that to prioritize the search indexing, which enables it to react very quickly to trending topics while still maintaining the ongoing drive to build a bigger network.
While Faroo are not yet making revenue off their search engine they have a plan that will allow them to share the revenue with both the users who are contributing most to the network and a number of charitable causes. It will be interesting to see if this will scale and be able to remain relevant – ReadWriteWeb asked this question and others, but they, and much of the commentary seem quite positive on the ability for Faroo to do well.
The other technology that requires a fairly high investment in infrastructure is remote storage. Services like Microsoft’s SkyDrive obviously provide robust and reliable storage but at a cost that most startups would find hard to compete against. Wuala (another strange name, and from Switzerland this time) have taken the same concepts of peer to peer networking that underlie the storage requirements of Faroo and similar projects and use it to launch a scalable file storage and sharing platform in a way that enables them to minimize their bandwidth and storage costs. They do provide a copy of the data on their servers but by default they serve it from the peer network first and only refer back to their infrastructure if there isn’t a viable instance available.
To leverage the network effect best Wuala reward active, online participants who contribute storage to the pool (with a reliable, high availability connection earning you a higher reward) with additional storage options. Their business model doesn’t rely on serving advertisements (which seems to be the de facto Web 2.0 pitch) but allows users to also buy additional guaranteed availability storage without having to provide capacity to the network in return. In the short time since they launched they’ve gained quite an active user base.
While BitTorrent may be the first thing to spring to mind when you talk about p2p and video today, that not may be true in the future if the new LiveStation platform takes off (and having used it for a while I can see why it should!)
Using technology licensed from Microsoft Research, and Silverlight to deliver the user experience LiveStation uses a peer to peer network to deliver a scalable live video broadcast platform. In a traditional experience the more users viewing the content the more infrastructure was required (either for the broadcaster or the content distribution network) but with LiveStation the opposite is true – the more users the better the platform is able to distribute the loan between peers giving a better end user experience without having to worry about scalability of the backend solution. As long as there’s enough infrastructure to seed the network they should be able to cope with any number of viewers – and use their dedicated infrastructure to insert adverts or manage other aspects of ensuring the service is commercially viable.
Is Peer to Peer the answer?
It’s hard to know at the moment how well any of these services will take off.
BOINC and it’s predecessors appealed heavily to the more technical end of the spectrum so installing and configuring a download wasn’t too much of a barrier to entry (though as it becomes more mainstream and they continue to not evolve the user experience to that audience I wonder if that will harm their growth). Both Faroo and Wuala are aiming firmly at the typical end user though so they’re going to have to overcome trust issues to get their engine running on as many machines as possible as quickly as possible, and they’ll have to make the experience so easy that after the initial installation the nodes work as effectively as possible.
At first glance Faroo has done a good job of making a fairly simple installation experience with sensible defaults to provide maximum benefit to the user and the network without too much downside (as it’s developed in .Net on a Windows machine there’s no need for additional overhead such as the Java runtime) but the Wuala experience has a little way to go – not only does it require Java the configuration process and usage isn’t as intuitive – but both of these are in beta so I’d expect them to improve and evolve over the coming releases.
LiveStation seems to have a great platform and user experience but needs more varied content to make it a viable long term winner – partnerships with more networks and perhaps providers such as Netflix or Blockbuster would make them a winner.
I think in the foreseeable future the traditional model of centrally hosted and managed solutions are going to continue apace but the viability of peer to peer solutions is going to allow more new competitors to enter the landscape and scale quickly and reliably. Perhaps we’ll even see a generic platform like BOINC evolve to provide a common services layer that handles the communication and security (and the initial install issue) and subsequent solutions simply deploy as add-ins with their own user interface, secure storage and networking / CPU asks of the user…
What those solutions look like… online time – and imagination – will tell…