thank you @Webb.Peter - I'll do my best to address your questions.
The main problem is there ARE NO OPEN SOURCE designs (that I know of) so collaborating with existing companies is near impossible, unfortunately, this is exactly what we aim to influence to change. If you know of some open designs, please let me know. ALSO - We must be very careful in releasing designs in this community, make sure everything is marked creative commons or lots of folks are just waiting to take them and protect them (trust me)
On containers: We are using reefers too!! (8'x10'x40') So thats great. I also want to be clear that we are NO WHERE NEAR a final solution, at all. Our design approach at this point is very simple, as modular and as mutable as possible. To that effect we have reserved approx 4 ft in the front for "elec/dry storage" that we have filled with power supplies that are meterable and controllable and the rest filled with 2'x4'x10' rolling racks (like bakers racks) on either side of a 2.5' aisle. Designing specific, fixed, customized racks at this point - in my opinion - is a waste of time, before we have set base lines of performance. WE NEED TO BE ABLE TO ITERATE QUICKLY without a ton of cost, this is already a very capital intense endeavor. What is more important is that our environments are controlled within similar bands of error and that we are using the same data collection methods. On that, our backend should be available soon, right @gordonb @DougieFresh ? also the "brain boxes" as we are calling them (pi+arduino+shields+sensors) should be as well, if they are not already.
Documentation is really hard to provide while innovating and designing. I think one thing I do poorly, and on accident, is share my passion and vision so clearly that it seems like we (MIT openag) have it all figured out. As a matter of fact, that could not be further from the truth.
We JUST landed 4 containers at MIT, they have not even been fitted out, its in process now and will be operational Jan/Feb. Happy to share renders of that.
We HAVE run a prototype (the clear/polycarb farm we have shown pictures of) that is 6'x8'x30' with a unistrut frame with modular 2'x4' shallow water culture and aero "component" trays. Happy to share details of that (pretty sure we have?) in pictures and maybe we need a video/hangout meetup?? anyone want to coordinate this? (@adrianlu ? )
I ABSOLUTELY agree you have insight to provide, that is why we built this site and it is my greatest hope. Maybe a new thread titled "server farm" with updates from everyone working on that?
I would say be very careful making any claims on the productivity / economic use case for the "server farm" it is still very early. Ive found that gaining corporate attention by saying "lets collaborate together" on this project and NOT "I have a solution for you" that drives a much more meaningful (and larger financial support) engagement.
I hope this helps and gets us moving forward.