Some comments from experience
Dear Trevor, I have some first hand experience in something similar :-) albeit in Europe and not Canada, but most of the economics remains the same.
I would say for a start that a total of 3 man/month of development of even a very good developer are not enough to come up with a product. Even a bare scaffolding+glue+cat pictures form of a product.
There are two ways:
1) you are assembling lots of things together and write some glue code and nice interface. This is what many vendors do with ZFS, for example; tuning, optimizing and painting a nice interface on top. I have looked at quite a few, and they all require at least 10 to 15 MM to be replicated (in a way that is functionally stable or at least that doesn't lose much data...)
2) you are making a new storage systems, that would probably fall in the "distributed" category. Even the smallest and simpler ones I know have a COCOMO estimate of several man-years of work (Ceph would require more than 200 man years, as a reference point).And would require at least a bit of research- distributed is a very complex field.
I would say that a real minimum would be more in the 1 or 2M$ range, including some expenses for things like promotion at events, marketing efforts and so on.
It took us 2 years with 4 people full-time to reach a MVP; thankfully we were financially supported by the European Commission with several research grants.
Can't wait to read part 2, though :-)