Toronto’s OpenCola lives on in Swarmcast

(cross-posted from my Globe and Mail blog)

Given the kind of publicity that Bram Cohen and BitTorrent have recently gotten by announcing deals with Warner Brothers and other movie studios and content owners, it’s worth noting that a Toronto-based company got there long before BitTorrent — at least in terms of the technology, if not the public awareness.

OpenCola was a technology startup with a peer-to-peer (P2P) application of the same name that was similar to Kazaa and Limewire, but years before either of those would become household names. On Thursday, the successor to OpenCola announced a $5-million financing deal with two Japanese venture-capital funds.

In the late 1990s, a young programmer named Justin Chapweske had developed a technology he called Swarmcast that was designed for sending large files over the Internet. It did this by chopping each file up into tiny bits, distributing those bits to many different hosts, and then allowing users to download those bits from any location — at which point the software would reassemble them into the original file. In other words, exactly the same process used by BitTorrent, except that Swarmcast was developed in 1999 and BitTorrent didn’t appear until 2002.


Justin wound up joining Toronto-based OpenCola — whose founders included Cory Doctorow, one of the editors behind BoingBoing. Unfortunately, that coincided with a downturn in technology markets, and OpenCola did not survive. The technology was sold to Waterloo, Ont.-based Open Text Corp. Justin Chapweske later started another company called Onion Networks and eventually reacquired the rights to the Swarmcast technology.

So why is BitTorrent a relatively well-known name and Swarmcast is not? Because the two took different approaches to commercializing their software. Bram Cohen chose the “open source” route and released the code for his software so that anyone could use or distribute it (so long as they didn’t charge money for it or claim it as their own). It quickly became the technology of choice for downloading everything from cracked software and illegally copied movies to pornography, although it was also used for distributing large files such as the various flavours of the Linux operating system. And that in turn got the attention of content owners.

Swarmcast, meanwhile, decided to focus on working behind the scenes with companies that would have an interest in distributing large amounts of content over the Internet — including distributing digital films to movie theatres. The company also helps power, the major-league baseball service, which distributes huge quantities of video and audio to baseball fans. Not as sexy as doing deals with Hollywood movie studios, but not bad either.

6 thoughts on “Toronto’s OpenCola lives on in Swarmcast

  1. Pingback: Main Page :: Global Nerdy

  2. Pingback: iBLOGthere4iM

  3. Swarmcast was open source and not proprietary. Swarmcast was not developed in Toronto, but was purchased by Opencola, which never did anything significant with it other than sink it. I believe you are also wrong about Justin re-purchasing the IP.

  4. Opencola devoted a lot of resources to the development of swarmcast. Lots of developers worked on it for many months.

  5. as one of those OC swarmcast developers, I can also tell you that Swarmcast was only superficially like bittorrent; for one thing, Swarmcast worked 😉 but the real innovation still unmatched was the insightfull use of the Foreward Error Correction — this is the same method used by cellphone networks and this technology (not invented by Justin, just adapted to filesharing by him) is the “free software” core of the otherwise proprietary swarmcast technology (FEC is still on sourceforge somewhere).

    Another SC filetrading innovation in the use of hash-trees allowed swarmcast to be used for real-time live streaming broadcasts; this facility is still largely untapped by the media although I believe it was employed by CityTV’s FashionTelevision to distribute their archive broadcasts (long long before YouTube and Google would make streaming media mainstream.

    So why did bittorrent skyrocket after the death of napster but SC didn’t? My theory is two factors, not the least of which was that Justin just didn’t go to the right parties, but also the SC implementation was in Java, and java for a windows desktop was not a popular opinion in those days. Had the desktop client been in C++ or even Python like BT, it may have fared better. As for being opensource, remember, in those days neither of the top browsers was ‘free’ as in free speech, and the swarcast client was just as free as in free beer.

    Oh, one other thought: we intentionally made SC difficult for bootleggers to spread illegal copies; to distribute content, you had to own that content; that probably shot SC in the foot relative to BT more than all the other reasons combined.

    But don’t shed any tears for OC Swarmcast: upon the collapse of OC, Justin regained the ownership of his wares and founded, and they are still very much in business still profitably serving the industrial sector that OC would not consider 😉

Comments are closed.