Same place as you buy everything else - the interwebs. If you're lucky someone might of open sourced the design.
100k isn't "perfect" - because then you install a sense of favouritism, install an us/them divide - not the right sort of thinking, IMHO. I don't see how or why anything you list should inhibit science within itself - Space is big, it's unlikely everyone would want to be in the same place, at the same time. And ignorance can be cured, if they are willing. Further, consider how much of "scientific discovery" was actually a mistake. Sometimes the wrong answer to that question is the right answer to another. Misunderstanding a principle can give rise to a completely new direction of thinking. Getting things wrong can teach far more than getting things right, it's how most learn the most effectively.
More ships and more stations are a defacto certainty IMHO - given the right timeframe - but the mass for dealing with 100k of population... Megatonnes, hundreds of. And that's before you start building to my spec - things like five meter thick NiFe skinned in 6" of titanium with a ½ meter thick MgAl foam exterior as radiation shielding and MMOD/strike defense.
To sensibly attribute for this, in a reasonable timeframe (I math about 3000 years just to lift the thermal dissipation hardware using NASA's new SLS system comming online soon) then I am of the opinion a combined deep space mining and industrial capacity could have the raw materials in near Earth orbit by 2040 or so(if we started by 2025) and large-scale residential facilities can begin assembly of components by about 2060. I personally don't see any other options, they all take too long.
Using exponential mining methods are pretty much the only way to achieve this, and at this point why just build one? Redundancy is not something that should be easily overlooked. There are distinct advantages to building more than one - response time for any support required being one, enough in the same orbital belt the right distance out evenly spaced could get from one t'other in a reasonably short time, and in the worst case of failures causing abandonment, the load can split between the two closest(initially). Having somewhere to go in an emergency is something that cannot be overlooked. There are many good reasons to build more, before we even start loading people.
A remote attack on such devices as guidance, communications etc would be feasible if we was to use COTS hardware in stupid places. But knowing this is already backdoored, we should sensibly develop our own, lacking unappreciable "features". TBH, comms shouldn't matter because we should be mostly automated - and with the distances we're likely to achieve light is just just too slow to consider as a transport medium(even surface » orbit has noticable delay) and we will need to devel some enhanced system - possibly based around quantum entanglement(which should be difficult to interrupt, or intercept).
Open source is already a solid thing. More and more companies are contributing to it, including names you might not expect, like Microsoft, Ford, Rennault, NASA - There's a lot of games that are about to have a vast adjustment to the rulebooks. The ones that can see this have begun adapting now. With regards to copywrite infringments, once I've looked at a design, that's filed. If it's assembled that procedure runs through my head backwards. After that, it's a little difficult to prevent it's further use - however, I'd never use it in anything anyone will get to see, and in any official capacity I'm unable to be seen to condone such behaviours.