We were reading Jinesh’s parallel computing with Matlab post at the Amazon Web Services Blog and some of the user comments got us thinking about Chris Anderson’s book Free. Justin and Eric both commented that the concept is cool (parallel Matlab instances on EC2) but quite pricey given MathWork’s per-CPU business model. This is true – Amazon’s cloud opens doors, but old school business models are trying to keep them shut. (Here’s a hint – if you need to login or contact a sales rep for software pricing, you are old school and friction-ful, and new school friction free business models will leave you behind).
Monkey Analytics is based on open source software (ie, free) and so we can offer our subscribers inexpensive scaling for their Matlab, Python, and R computational needs. Each additional CPU is not an additional pricey MathWorks license, it is just another incremental CPU hour for the subscriber.
Our value add is not in the licenses themselves (those open source software projects all exist in their own right) but in the way that we package these systems for our users and subscribers to enjoy. MathWorks has to do everything and charge for it – their 2007 revenue was $400M. We plan to be a “$0 billion dollar business” – imploding these license fees as we make computational analysis far more affordable to more people. Stay tuned to see how we do!
Update: the article notes that a MathWorks representative is willing to discuss use cases with people who wish to do something like this. At least they’re thinking about it, but the core point remains – they are still competing in a world of free. (And on this note – we would happily resell Matlab as a Monkey Analytics engine if the pricing was usage based and reasonable).