At the end of October one of the admins of the world's largest Bittorrent site sat down for an interview and predicted the protocol's demise. Citing Bittorrent, Inc.'s corporate ties and some technical limitations, brokep announced that The Pirate Bay was working on a new protocol to succeed Bram Cohen's Bittorrent. The idea's been percolating throughout the filesharing scene since then: a small survey of top site admins conducted by TorrentFreak found opinion divided over whether Bittorrent will be replaced.
It's true that the protocol's been asked to do things that its
creator didn't envision. Clients now use encryption to get around ISP
traffic shaping and sometimes pad
files to improve interoperability with other networks. DHT
functionality, which removes the need for a central tracker, was
implemented in a chaotic, piecemeal fashion. Private trackers have to
monkey around with torrents' announce URLs in order to monitor
individual users' activity. Torrent files lack metadata. Traversing
firewalls remains an issue. And various researchers have created custom clients that
prove the protocol can be subverted by selfish users. There are
tacked-on, vulnerable and subpar aspects to the way Bittorrent works
— plenty of room for improvement, in other words.
But assuming a technically superior standard is produced, will it be adopted? It's easy to find examples for and against: the Ogg Vorbis audio codec offers better sound quality than MP3, no licensing entanglements, and several awfully-cool features (like the ability to reduce a file's size without reencoding it). But Ogg has never really caught on. Some users employ the also-technically-superior WMA and AAC formats, but only to the extent that Microsoft and Apple force them. For most users, MP3 seems to be good enough. On the other hand, online video has adopted new codecs almost as soon as they become available, moving from VCD to SVCD to MPEG to DivX to Xvid and beyond. The situation's so complex that utilities exist for the sole purpose of untangling a given AVI's miasma of codecs.
What makes these cases differ? It all comes down to timing: consumers will switch technical standards so long as doing so carries few costs (i.e. only requires that more free software be downloaded). Ogg Vorbis hadn't attracted enough attention by the time portable MP3 players arrived. Once the supply chain for MP3 decoding chips was established and a generation of compatible players purchased, the game was pretty well decided. By comparison, only a handful of exotic DVD players bother to support the video formats commonly found on P2P networks. Most portable digital video players still count on users recompressing their files to save space and conserve CPU cycles. Once there's an established infrastructure — of either hardware, accumulated code or simple corporate momentum — consumers may stick with suboptimal technical standards. But prior to that point, users will stay close to the cutting edge.
Bittorrent seems to be on the cusp of this transition. Some hardware devices are coming to market with the standard baked in, but not too many. Various organizations like Miro, Joost and Blizzard Software are building parts of their business around the protocol, but not in an irreversible manner. If Bittorrent gains much more attention, its supporting infrastructure of trackers and open source projects will likely trump whatever advantages a new standard can offer. But I think that there remains a window of opportunity for elite users to popularize a new protocol, should they settle on one. Brokep and his peers still have a few months to steal BT's thunder.