Hello David!
So I had a crazy idea of renting a windows dedicated server to serve as a winflector server to try to solve the web browser problem.
...
Anyways, anyone interested or is this a dead end idea?
...
So here is my response while wearing the Corporate IT Hat:
1) SECURITYBefore anyone signs up for this kind of a service you need to very precisely spell out how the security aspect of MY data residing on YOUR server is going to play out. By that statement I do NOT just mean "is my data secure?" as if it were in it's own cloud hosted server...you need to account for the fact that multiple user IDs would now exist on the same box and therefore you'd also need to account for how that segregation is to be accomplished.
2) PLATFORM#1 above assumes a single server, is that indeed what you are thinking? How about looking at any of the cloud platform services, purchasing compute & storage credits which are then applied to UNIQUE VMs that host user specific environments?
You see, this approach, while I suspect more costly due to underlying OS licensing requirements, would certainly account for a lot of the SECURITY worries. After all, each and every time I connect to run one of these remote apps i'm essentially logging into my very own hosted VM environment.
3) PERFORMANCEThere is a big difference in implementing this within your LAN vs such remote configuration as what you propose. I would want to prove this out first so the sort of usage people are expecting to see (say YouTube access) can in fact be facilitatted.
Regardless of whichever approach you pursue in #2 above, the speed at which that remote session is instantiated would play a big role in how "happy" the end user is. After all, if it takes me 1-2 mins to spin up a remote FF window, and even then I can't honestly use it to browse my favourite YouTube channels (just a use-case here LOL), then is it really worth it to me?
In other words, we should define real use-cases and spell out what the end-user expectations are going to be met by each one.
4) APPLICATIONSIs the intent to create a blank-slate remote OS platform which each user then adjusts to his/her liking? I think this is largely driven by the PLATFORM approach you decide to pursue. If we are all in a single virtual box, how do you know manage the application configuration?
Are we all permitted to install our own Apps? Oh, but you need Admin access for that?...sooo, do I get that Admin access (hell NO, b/c then you open up everyone else's data to my inspection), or do you exclusively own the Admin role and work through the AppInstall queue requests???
OK, now from the end-user perspective: I do think this approach could work if you were to really scope this down and perhaps just establish a package of 1-3 Apps that are accessible through such a service. Would this still be enough to entice folks to pay $15/month? Hard to tell...
Corollary: this proposal is a perfect example of whee an established business (AN) could step-in and put their weight behind this. Given my experience in the corporate cloud procurement space I am comfortable saying that economies of scale apply and based on who the licensing request comes from will dictate how that conversation plays out.