16 Dec 2003
The other day, Steve Gillmor wrote about BitTorrent and RSS and how they could be combined to create a “disruptive revolution.” He’s half right. RSS and BT are indeed two great tastes that taste great together, but Gillmor’s vision is upside down: we shouldn’t use BitTorrent to carry RSS, we should use RSS to carry BitTorrent. Let me explain.
RSS (RDF Site Summary) is a simple format for syndicating content on the web. These days, the most common application of RSS is subscribing to weblogs: you tell your computer to check an RSS file for changes every so often, and then it notifies you when there’s something new to read. If you’re like me and you read one metric shitload of news every day, this is a life-saver.
BitTorrent, the brainchild of Bram Cohen, is the current cool-kids’ P2P program. It works sort of like Kazaa, but at a lower level. It doesn’t handle searching for new files, it doesn’t have a media player, it just concentrates on downloading big files efficiently.
Okay. Two solutions in search of a problem. Here’s a problem:
I am addicted to the show Alias. I watched the first couple episodes of season two as it aired, and I was hooked. In my honest moments, I’ll admit that the show’s appeal is mostly due to the callipygian Jennifer Garner. It’s a weakness; we deal.
But it gets worse. I go out on Sunday nights, when Alias airs, and I don’t want to give that up. That’s why God created the VCR, I know, but to compound the problem, I don’t have TV. I don’t want to have TV, because I love the feeling of superiority that I get by not having it.
This system is at tension, it has no rest, its forces are unbalanced, it wants to be resolved.
The internet, it turns out, is great at resolving different kinds of tensions, and this is one of them. After a few weeks of missed episodes, I realized that with a little patience, a P2P program like Kazaa was able to fetch back-episodes with aplomb. Each file is around 450 megs, fairly high-quality video, with commercials cut out. I start a few episodes downloading, and by the next evening, they’re ready to watch, whenever I have the time.
After a few weeks of enjoying this, a new tension emerged: I had caught up with all of the old episodes, and I had to wait a week for each new one. The problem is that the Kazaa protocol isn’t especially well-tuned for getting brand new files: first someone has to record the show as it airs, cut out the commercials, and compress it to a reasonable size, then seed it on the network. Then, it has to slowly propagate to its peers, each transfer taking hours. It might take three days before it’s available on enough peers that I’m able to even find it, let alone download it.
The solution is BitTorrent. BitTorrent operates on similar principles to Kazaa, but it’s tuned differently: it excels at downloading files that are new or currently in high demand. It breaks large files into many small chunks, and coordinates their assemblage, so that users can tap into a swarm and distribute the load evenly. At the same time that you’re downloading a chunk, another user is downloading an earlier chunk from you — no one server is overwhelmed, and the more popular a file, the higher its availability is. It’s perfect for large files that are most interesting when they’re fresh — in other words, it’s perfect for TV shows.
In many cases, I have been able to use BitTorrent to completely download a new TV show mere hours after the show airs. Like a TiVo user, I’m no longer bound to a specific time to watch my shows. I’m free to go out on Sunday night and still watch my show while it’s brand new. TV is now asynchronous.
Life is good.
RSS has revolutionized the way I consume some types of content, like weblogs and news. Instead of compulsively reloading a huge list of bookmarks looking for updates, my RSS reader scans them all every hour, and summarizes what’s new — allowing me to keep current on more sources with far less effort and waste.
Imagine if I had the same efficiency retrieving video files. With the addition of RSS, BitTorrent could really be taken to the next level, and I’d be able to forget about the plumbing of TV altogether. I want RSS feeds of BitTorrent files. A script would periodically check the feed for new items, and use them to start the download. Then, I could find a trusted publisher of an Alias RSS feed, and “subscribe” to all new episodes of the show, which would then start downloading automatically — like the “season pass” feature of the TiVo.
Illegitimate uses of this system would obviously abound. But the potential legitimate uses are huge as well. For one, traditional content providers (like the TV networks) could take advantage of the demand for their programming by scooping the copyright infringers. If ABC released Alias on BitTorrent with advertising built in, the file could be delivered to their audience very fast, and would cost them next to nothing in distribution costs. The economics of producing video programming would be upended — each viewer of the program would, in effect, foot the bill for a tiny slice of the distribution overhead, causing a massive component of traditional media company infrastructure to become obsolete.
The point that Gillmor misses is that RSS files and JPEG graphics are small, and BitTorrent’s overhead makes it ill-suited for small files. The BT/RSS combo is potent, but its potential lies in using each technology’s strengths — RSS for syndication: low-overhead references to fresh content — and BitTorrent for the heavy lifting: swarming transport for high-demand, big files.
The beauty is that this synergy can so easily be realized. You can start publishing RSS feeds of BT files today: just point to a
.torrent file in the
<link> element. A user who consumes the feed in their favorite RSS reader will be able to start the BitTorrent download with one click. Then, as the creators of RSS software catch on, they can automatically start the BT program when they encounter a new
The result: the TV distribution networks are completely end-run by an ad-hoc, decentralized, loosely-coupled network. And in the process, significant opportunities are afforded to independent content producers of audio and video to reach a mass audience with insignificant distribution costs.