Correct me if I'm wrong here, but isn't the failure of smart homes attributed to the SaaS model itself. Nest devices had to be killed because Google wouldn't support the subscription anymore. What it needs is licensing model and not a subscription. Get a smart home installed on your home like you'd install Windows on a PC. We are basically wanting to install a server in every home to cut off recurring costs on vendors and even dependency on a single vendor for that matter.

The bottleneck here could be the network interface once you are away from your home. That is where standards struggle.

Looking at the sibling I currently see right below hits the nail on the head as in vendor lock in is bad for both SaaS and perpetual licensing models.

I think none of the models currently work well for the consumer. If your SaaS provider shuts down its service you have to replace all the hardware too because of vendor lock in.

If your perpetual license vendor stops making this produ t line or changes its proprietary software incompatible say 3 years down the line (which from a software development standpoint is reasonable actually) you can't buy another light switch for the new room you added by subdividing the basement "play room" that would work with your original installation. Software incompatible. New software or other vendors' software won't work on your 'ancient' hardware.

Vendor lock in bad.

I want a smart home that is isolated. Given that i know all the ways a computer system can fail, I want my smart home to function even when Amazon/Google/Microsoft cloud goes down. Also , I dont want my smart devices to send or receive ANY data from external servers. Of course that's incompatible with current business models. So I prefer my home to stay dumb.

In principle, the answer is getting together and standardizing on generic protocols.

Unfortunately, there are multiple reasons why this can't happen in the current environment.

The business reason is that all the big tech giants are now deeply accustomed to lockin-as-a-model, and you can't have lockin if you have a standard that everyone can use equally. So, literally everyone in the space sees it as their incentive to let other players conform to their standards, but not to conform to anyone else's standards. This does not result in a standard, de facto or de jure.

(I made the same objection to the "metaverse" concept a couple of weeks ago. It can't work as some people are envisioning; the business proposition at this point is complete nonsense. Every participating company will be happy for other companies to voluntarily walk into their walled garden, but no company is going to walk into somebody else's walled garden.)

The technical reason is that any attempt to standardize at this point is more likely to produce a mess like bluetooth than any sort of coherent standard. Plus you get an echoing of the business model problem where every participant will see the standard as a weapon to be fired at their competitor rather than something to conform to for mutual advantage.

Not helping any of this is the problem that the problems that they are solving just aren't that compelling. This limits the amount of value they can extract from me by virtue of the fact I'm not paying for more than I'm getting from the 'smart home', which doesn't really leave them with much money to play with. It also doesn't help that the camel's nose is already under the tent in terms of not simply enabling the user but trying to control the user, with things like letting this authority twiddle with my air conditioner and that authority twiddle with my hot water heater, and in 2021, I have less than no confidence that buying into a smart home won't continue this trend.

Personally, my assessment is that the tech industry of the 2010s and the 2020s is structurally and temperamentally incapable of producing a high-quality smart home. The best they can do is create a locked-in walled garden that works great if you buy into it from top to bottom (and pay the subscription fees!), but gets abandoned in three to five years and then becomes totally nonfunctional as they pursue the next thing.

There are some standards already and a lot of the big companies use them already. Amazon's Alexa uses Zigbee to communicate with 3rd party devices and can act as a defacto hub. Zigbee is an open protocol with no licensing fees (or at least cheaper than Z-wave, it's slight more reliable, but more expensive, cousin protocol).

Many devices are now opting for control over HTTP which is fairly reliable, even if your outside internet connection goes down. It's also easy to setup and discover. There is a larger security concern for these types of devices, but common sense will go a long way towards keeping you protected until more rigorous standards can be agreed upon.

Home automation does not, and in my opinion, should not require and outgoing internet connection. My network is accessible from outside because I want it to be and I've accepted the potential risks (though tunneling into my own network is always an option should I want to go that route).