As stated in other posts, NASA will have a need for a LEO post. LEO is cheap to actually train astronauts compared to BEO. The radiation environment is milder, thus they can make a few tours of duty without tripping over their life radiation exposure limits. And emergencies can be handled more easily (like in hours time, rather than days or months). But above all, a lot of technologies that need to mature can be done in LEO safely until they are reliable enough for BEO.
Thus, how could a Commercial service help for all this research? Well, there should be a double approach. First, is the actual human experiments and technologies (like suitports, ECLSS, radiation mitigation drugs, etc.) and the second are straight experiments.
I'm not sure that a single station could cover it all. There might be some ways to leverage a single point. But it would have to have certain limitations. I still believe that microgravity and radiation experiments might be better served by a free flier. Nothing prevent such free flier to actually launch and return to the station. But the microgravity quality, the simplification of being able to actually conduct dangerous experiments, and the possibility to gain altitude to increase radiation exposure without endangering the crew, but having the possibility of actually having the crew setup experiments and process samples for return, are all clear strong points of such proposal.
The services that a LEO station could supply to a module would be many, and greatly simplify the BEO effort to just those technologies that need maturation. Let's go over the main HED need that need LEO for development:
* Radiation shielding during the cruise and surface-sorties phases of a Mars landing mission
* A new environmental control and life-support system.
* Habitats for crews journeying to and living on Mars.
* Extravehicular activity suits, otherwise known as space suits.
* Crew health. There are medical unknowns associated with long-duration spaceflight, (and Mars surface activities).
First we have to assume that a commercial station can do certain things cheaper than NASA. I'll take that as an assumption and won't go into discussing it. That's material for another thread.
Let's remember that communications, thermal environment, MMOD, molecular oxygen and power supply are different problems for LEO and BEO. Also mass is a lot cheaper and can be done with off the shelf LV. Thus, replicating any of those things wouldn't be a useful investment. Radiation is milder and might have difference on high energy particles in LEO. But as a milder first approach is the right middle step to take.
Second, if they can get some extra users (i.e. non-NASA), they can lower costs. Thus, my proposal is that the commercial station would supply all this services:
-Crew transport (including emergency evacuation vehicle).
-Cargo transport (pressurized, unpressurized, return and dispose).
-Power generation, storage and distribution.
-Continuous communication with ground stations (both telemetry, voice, video and wideband)
-A docking/berthing port.
-A proven ECLSS.
-A traditional EVA port.
-Might supply some fluids, too.
-Confidential space.
Thus, NASA could make a custom module, dock it with the commercial station. Inside that module, they could have their suitport, the new ECLSS, test long duration screws, try delayed communication protocols, train and improve on suitport EVA procedures, etc.
Yet, they would be just a closed port away from a safe heaven. And if they had any contingency on the suit ports, they could have the traditional EVA port (plus some suited astronauts ready for rescue). They wouldn't have to worry about power, communications, supplier or even waste disposal (unless they want to simulate that).
The rest of the commercial station could have "green" NASA astronauts, some on support, the commercial contractors for general experiments, and tourists. The latter part would have some issues on control access. But if the station is big enough and NASA accepts international space agencies nationals on the "LEO side" of the station, with the required institutional agreements, it shouldn't be much of a problem.
In fact, the concept of a permanently closed door that can only be opened on emergencies might work well enough. They could share general quarters and galley. And they could use a double door activated by the ground control for accessing the NASA or Commercial clearance-only parts of the station.
One interesting part is that NASA might send radiation exposure experiments on the the commercial side, with no astronaut ever touching it. And they might gain access to the free flier platform for higher radiation exposure and dangerous things (like validating fire suppression systems). Not to mention that they could contract for the commercial operators and agree with the foreign nationals (and even tourists) to run certain physiological tests, thus increasing significantly the sample size.
This would allow to concentrate budget and efforts on the BEO specific issues. And who knows, the closed-loop ECLSS might even be interesting for the commercial operator, whom might license it back. Might even develop a commercial version for itself that might be usable in the future for BEO.
I find this model of services very intriguing. If reusability can actually be used economically for LEO. There's potential for a significant reduction on costs. And if not, having the commercial suppliers apply creative thinking and ingenuity to supply a given service for less costs, should help significantly.
The truth is that this would require a very serious effort on actually specifying the services requested. And developing a certification process for it. I firmly believe that if the ISS can be extended to 2028 and commercialization is extensively used, by that timeframe a commercial replacement will require little new development. By 2018 crew and cargo services will be characterized well enough. Nanoracks and CASIS will have standardized experiments significantly. And I understand that imaging and communications will have a pretty good degree of development.
I understand that some detail might have to be traded. As in free flier or not. Or external arm self relocating arm or simpler versions. Or docking vs berthing. And a long list of etc. But clearly that's where the commercial suppliers can actually lower the cost. This would not be about the engineering exercise but actually closing a business case.