Hearing about the PS3 Folding@Home Project got me to thinking about distributed computing. Sony is using idle cycles on PS3 machines to do medical research. This is similar to SETI@Home, a desktop screen saver that searches for extraterrestrial intelligence. (The Folding@Home Project can be run on a desktop too, I have it running in the background currently.)

  Building on the distributed screen saver concept I have a multi-phased idea:

  1. Build a very slick screen saver to be given away free.
    The amount of installs must be very high so people should feel they are getting a high-end screensaver for nothing (for now).
  2. Build a distributed platform into the screen saver so that you can sell off compute time on any machine that is running the screen saver.
    If #1 worked well, we should have a compelling distributed platform, then we have to get the compute sales up.
  3. To keep the pipeline going and keep people interested we start giving money to the users based on how much computing their computer has done. This will encourage them to leave the screensaver installed and running (perhaps even running in the background during use throughout the day). We’re basically paying them for their spare computer cycles.
  4. At some point we could have the option of disabling the distributed portion and just selling the screen saver. Of course that wouldn’t be optimal because we want a constant stream of income.

So, in the tradition of this blog. Reasons this idea won’t work:

  1. It sounds hard. 🙂
  2. Creating a compelling Screen Saver is easier said than done.
  3. Limited market – I have no idea how big the market for distributed computing is, but not many people need the immense computing power that distributed computing provides.
  4. Limited computations – Distributed computing is limited to attacking problems that can take advantage of massively parallel processing.
  5. Reduced costs of supercomputing.
  6. The type of people who are likely to install screen savers for this type of effort are likely already using Folding@home or SETI@Home. Getting them to switch would be tough, particularly if you don’t start paying from the beginning. Of course, if you’re going to start paying from the beginning you’ll need investors (who disregard the previous 5 points :))