As anyone who has ever tried to distribute a large amount of data over the Internet knows, bandwidth while being plentiful is not really all that cheap. Pricing for outgoing bandwidth can vary considerably, but is never less than about 25 cents per GigaByte (GB), and can be as high as eight dollars!
For the small, independent musician, film maker or software publisher, distribution of their materials can be the most expensive part of their business. This is particularly problematic if the material being distributed is not being paid for, such as in the case of Open Source software, or free promotional content from an artist.
Centralized distribution methods on the Internet also have the problem of massive congestion when new, highly sought material becomes available. For example, whenever a new Linux Kernel becomes available at Kernel.org, the site’s 1 Gigabit (Gb) pipe quickly becomes saturated and downloads can take hours. And forget it when a new Fedora Core is released — downloads of the ISOs can take days, if you are even able to connect.
Fortunately, there’s a better way: Peer-to-Peer (P2P) distribution. The idea is that if everyone who wants a particular file work together to redistribute it amongst themselves, then the load on the central server will be much lighter. In the ideal scenario, the server would only send out the file once — different pieces to different clients — which are then distributed among all those who happen to want it.
There are many examples of this type of technology, including the now defunct Napster, and the newer generation Kazaa, eDonkey and Gnutella. Unlike Napster, which was killed by the Recording Industry Association of America (RIAA) by way of a copyright infringement lawsuit, the new systems are completely distributed — as in, there is no centralized server keeping track of what files are available from the clients — they create their own “universes”, if you will.
These systems are best known for illegal sharing of music, and more recently, TV shows and movies. The inability for the RIAA and the Motion Picture Association of America (MPAA) to sue a single site or company is causing great consternation to these organizations, who are in desperation polluting the universes with corrupt files (or, Madonna using some very colorful language), and suing individual users.
Another, in many ways better, option is a system called BitTorrent. BitTorrent doesn’t create an independent universe, but instead relies on a web-page to make available to a client a small file which gives details on where to contact the Tracker for the large file to be downloaded.
The Tracker is a small server which simply keeps track of which BitTorrent clients are connected for a particular file. The Tracker itself does not know anything about the file, and does not have a copy itself. One can run their own Tracker, or can use one which is publicly available (there are many).
The last step in the file distribution process is that there must be at least one Seeder, which is simply a BitTorrent client with a complete copy of the file, and is generally started by the person wishing to distribute the file (and, hopefully, is the copyright holder of the contents). Clients wishing to download the file first contact the Tracker, and then all the other clients currently downloading the file.
At first all the clients will be downloading from the Seeder, but as it sends out different parts of the file to each client, these quickly start exchanging data directly with each other, trading data which they do have for data they don’t. Thus, the Seeder may only need to send out a total of one copy of the data, regardless of how many clients are downloading.
In fact, the amazing thing about this design is that download speeds will actually increase for each individual client as more clients are downloading. And, because the BitTorrent system relies on a web-site to begin the process, it has no searching ability itself and thus does not have any bandwidth overhead associated with the searches. Search bandwidth is a large amount of the overhead in the other systems.
And another key point here is that because BitTorrent doesn’t try to hide like the others, it is generally used more for legitimate distribution of content — everything from Linux distribution CDs to music or video legally released for redistribution by enlightened artists or public agencies. See, for example, bt.etree.org or legaltorrents.com.
Not to say that BitTorrent can’t or isn’t used for piracy — it can and is — but of all the P2P systems available, BitTorrent is the one used the most legitimately. The author intentionally made it this way, knowing that the P2P systems were getting a bad name thanks to the efforts of groups like the RIAA and MPAA.
So, for anyone who’s wanted to experiment with a P2P system but was nervous of being sued, you’d be encouraged to visit bittorrent.com and download the system. And for anyone who needs to distribute large files to the public, there’s no better option. Please, though: make sure you own the copyright — I’d hate to be accused of encouraging piracy.
Published in the Victoria Business Examiner.