For that price you could hire an hw hacker to find out and make it public.
With the MDK you'll have to sign a NDA and keep it as secret as nuclear launch keys.
I've got a similar problem on the SW side:
I'm currently evaluating the (arm-based) crio for an diagnostics application. The hw looks fine, but the SW side is - to be very polite - problematic.
We need the *full* sources from bootloader to kernel to userland (note we're not interested in the LV stuff at all) or the necessary documentation to write the drivers on our own.
(which we'd do anyways, as it seems NI bypassed the proper infrastructures like IIO and done most in userland - we need the ADCs via IIO)
As things stand now, NI only offeres some old patched-to-death vendor kernel far from mainline (okay, I'll handle that myself) and - tighten your intellectual seatbelt - an proprietary binary-only kernel module. This, obviously, is technically blunder - totally unusable for anything but childs toys.
Maybe I *could* get the necessary documentation by the MDK (*assuming* it also copes the controller side, not just the modules), but it requires an NDA, which would have to break, as my drivers will have to be GPL. So, unusable.
In the end - unless NI discloses their precious "trade secret" - we end up w/ two options:
a) reverse engineer everything and publish it on our own
b) drop NI the supplier list
Considering NI is already very expensive, I doub't my client will invest in the RE, but instead
find another supplier or design own hw.
OTOH, having all information, developing fresh drivers and bringing them to mainline would cost less than 10 units. So, NI could resolve the problem once and for all for a relatively little investement and then could be taken seriously in Linux community. Until then, their products are pretty useless for any serious Linux applications.
Linux Embedded / Kernel Hacker / BSP / Driver development / Systems engineering
