I did something similar for Wordpress/Woo CMS so right now I'm fully abstracted from the IaaS layer. That said that was way easier. I can only imagine if today consumers can lift and shift their laptop. I mean enterprise did it, mobile OS did it so why not consumer pcs?
I've taking a MVP approach so will just fulfil what I need to tier out my applications that requires lift and shift and offer extensibility to AWS. But you are spot on on using an axe to kill a chicken though. The solution will be pretty since it's enterprise grade but too heavy/costly for consumers.
Peeped into the VM WS community chats and looks has its fair share of issues too. Exploring free eSXi as well.
Demand and supply. There is very few of such enterprising requirements in the consumer space, and also, there is no good solution to it that comes cheap either. If there is no big $$$ in the consumer space for such requirements, there is no reason to dump in all the research and support into that area. Not surprising at all.
Enterprise space is norm because companies normally pay top $$$ for such kind of implementations and the business requirements and risk can stomach it.
Mobile wise is different. Today the common platforms are simply Android and iOS. Windows mobile is out, Blackberry is not in focus, Linux mobile OS is still an emerging odd ball. If your mobile space is as convoluted as the general desktop/laptop space, it will be a lot more difficult too. That being said, between one android to another, the changes are not all that big. If you are shifting from Android to iOS or vice versa, the migration wouldn't entirely smooth unless you stick with some common services. Mobile environments are still very well controlled, quite unlikely all the bells and whistles and different configuration you can have in the general computing arena.
If you are just treating AWS like yet another virtual hosting company, then what you are trying to do still works. However that is just a very narrow use case of AWS. Throw in a larger architecture that involves multiple systems and involves multiple managed services, the scenario will be quite different.
For example, I wouldn't be setting up my own database on any instances unless there is a clear advantage as oppose to using fully managed RDS. My own bespoke solution will choose S3 to persist long term data as oppose to storing them in DB, or files, or setup my own NAS/SAN solution. Neither will I be using my own email system if I can use SES or other email providers outside of AWS like mailchimp/mandrill. When architecting in manual cloud environments like Azure and AWS, I will make use of their scaling out architecture instead of static instance solutions. For those whom prefer dockers chooses ECS, prefer Kubernates choses EKS and so much more.
In the end, the one-to-one migration from VMs into cloud is not all that simple, and neither will you stop at it if you truly wants to embrace the cloud platform, right ?
For consumers, such techniques aren't attractive for there is little use cases for it. Most consumers will not even want to run their environments in the cloud remotely yet. That is like going back to thin clients with mainframes. Cloud initiative is definitely doing that, with introduction of solutions like Chrome OS, but Chrome OS isn't big time yet. Cloud providers offer solutions like AWS Workspaces, Azure WVD, the good old Citrix VDI, but all these are more for corporate consolidation to modernise the on-prem infrastructure, not a consumer space initiative.
Until we have some technology more compelling and homogenous to offer, I don't foresee consumer going back to the mainframe days yet. It may happen for general computing eventually. We have already observd the gaming industry is going this direction; Steam, Google Stadia, Geforce Now, Project xCloud are moving towards it.
So you see, for ease of use for consumer, it is not likely system migration. It's back to the good old mainframe days. Thin clients, thick servers in the cloud.
For your own knowledge, I think you can venture for your own use case, but I really doubt it is the eventual direction for everyone, since it is so much easier to just have a system that works in the cloud with high network bandwidth everywhere. I feel it is the direction for metropolis, while it will take quite awhile in the suburbs.