Google's parent company is disabling old smart home devices

[wkearney99]

Yeh, that's a completely bogus claim. The only thing that has to be done via cloud is making money via the cloud. And it does make it a bit easier to connect remotely, but of course any company could reasonably provide that feature without the rest of the cloudiness. Well, there might be things that require enormous processing, like reducing all of the gigabytes of data they've collected on you in order to sell it for money I guess.

 
You GREATLY underestimate the value of connecting remotely.  It's not about just being able to do something when you're not on site.  It's about multiple things (services, devices, phones, tablets, voice) being able to interact without everything HAVING to be on site or having to be CONFIGURED to allow connection into the site.   Being able to do NAT traversing and the like has been a huge hassle.  Making use of cloud resources for this takes  of a lot of the very annoying dirty work out of the equation.  
 
This is not without risks, costs and side-effects of third party motivations.  TANSTAAFL always applies.  But it DOES get to the heart of what consumers want... "just work" and "cheap".  Their existing crappy-ass home router stops becoming (as much of) a support problem.  Or their two-cups-and-wet-string ISP bandwidth, which is being neatly forced into improvement by virtue of cloud-based services like Netflix generating consumer pull.

Meanwhile, services like AWS and Bluemix, Xively (and no doubt countless others), coupled with SSL certs being more readily purchased by consumers means a lot of security hacks can be dealt with more effectively.  

 

[123]

At the risk of sounding like a shill for openHAB, even this open-source HA solution touts its own "cloud service": https://my.openhab.org
 
Supporting Bill's point, they introduced it to simplify remotely-access to one's in-home openHAB server. No router-tweaking required, no dynamic DNS, etc.

 

[Dean Roddey]

You GREATLY underestimate the value of connecting remotely.  It's not about just being able to do something when you're not on site.  It's about multiple things (services, devices, phones, tablets, voice) being able to interact without everything HAVING to be on site or having to be CONFIGURED to allow connection into the site.   Being able to do NAT traversing and the like has been a huge hassle.  Making use of cloud resources for this takes  of a lot of the very annoying dirty work out of the equation.  
 
This is not without risks, costs and side-effects of third party motivations.  TANSTAAFL always applies.  But it DOES get to the heart of what consumers want... "just work" and "cheap".  Their existing crappy-ass home router stops becoming (as much of) a support problem.  Or their two-cups-and-wet-string ISP bandwidth, which is being neatly forced into improvement by virtue of cloud-based services like Netflix generating consumer pull.

Meanwhile, services like AWS and Bluemix, Xively (and no doubt countless others), coupled with SSL certs being more readily purchased by consumers means a lot of security hacks can be dealt with more effectively.  

 
When I said they could provide a solution, I meant a server of their own that is purely for remote connection. So your automation system connects outwards to their server (securely) and your mobile device connects securely to the same server, and it acts as the middle man to allow them to set up a connection without any need for port forwarding and such.

 

[Dean Roddey]

At the risk of sounding like a shill for openHAB, even this open-source HA solution touts its own "cloud service": https://my.openhab.org
 
Supporting Bill's point, they introduced it to simplify remotely-access to one's in-home openHAB server. No router-tweaking required, no dynamic DNS, etc.

 
OK, yeh, so the same thing I was saying above. It's not really got nothing to do the 'thugh cloud' at all probably, if it's just to provide that trusted middle man feature for remote connections. 
 
But, it's one thing to do that when you are an open source product. If it's not available, no one paid you anything for it. If you are a commercial product, then it brings along a LOT more responsibilities and costs.

 

[rismoney]

Yeh, that's a completely bogus claim. The only thing that has to be done via cloud is making money via the cloud. And it does make it a bit easier to connect remotely, but of course any company could reasonably provide that feature without the rest of the cloudiness. Well, there might be things that require enormous processing, like reducing all of the gigabytes of data they've collected on you in order to sell it for money I guess.

 
You are a very cynical about the cloud and it is not surprising.  I have a lot of respect for you, your deep automation skills and accomplishments and your product.  However, cqc as noted throughout this thread I just read coincidentally- http://cocoontech.com/forums/topic/27882-what-is-wrong-with-cqc is what is steering me clear of your approach despite your commendable community request for improvement.  Rejecting standards, building proprietary language schemes and drivers, and anti-cloud are certainly not improving the state of things and not something I can get behind.   I am not saying Homeseer or Openhab have it right, but in my current search for home automation controllers cqc is unfortunately not making it to the table (and cost is not a consideration even at unobtainium prices (love that name))
 
Though not related per se to home automation, I do use github and aws services quite a bit for my work, and it's an eye opener into how to do cloud smart.  Web/Service hooks are the future and I recommend embracing them.  The winning companies are realizing great api's and simple integration work and i think that's why smaller companies attaching to them say (travis-ci or circle-ci as CI service companies for code testing) are succeeding.
 
In evaluating Homeseer, my biggest complaint thus far (aside from the product), is the sheer amount of threads on their forum that go unanswered, unsolved, and don't provide a conducive environment to create and share.  I read a thread saying that you will never see a $10 plugin as homeseer takes $9 or 30% from every plugin sale.  Today I am trying out the sonos plug-in w/ Homeseer, and the paradigm for all these devices wreaks awkwardness, (even if 80% functional).  It feels like square pegs and round holes, shims and plumbs.
 
Openhab is on deck, but the Java alone looks rediculous.   I am nervous about its configuration paradigm as it looks like it wreaks of endless tinkering.  I am happy some fine folks are adding upb to it here, https://github.com/openhab/openhab/pull/3883/files but even a minimalistic approach required nearly 1600 lines of code.  Home-Assistant is another one written in python iirc which seems simpler to approach and I do LOVE yaml.
 
Unfortunately the root of the problem is, unless you are a programmer yourself, to customize anything you will be reliant on others.  I probably will spend the $300 on hs3pro during bf in may sale, until the point where I roll over and write my own (probably around golang in containers, and a unique message bus architecture)

 

[Dean Roddey]

The thing is, this community provides little in the way of viable revenues for a company. Commercial automation is the only place to make any real money. No one in commercial automation is really interested in stitching together various third party bits and pieces. They want highly integrated products, where if something goes wrong they know at whom to point the finger (almost 100%, though of course there are exceptions because of the devices themselves and some environmental issues), and robustness and reliability and quality are at the top of the list of requirements. All of those things come from having a high level of control over the quality of the code, and that's only possible if we create it ourselves. It also means we can prevent people from shooting themselves in the foot to a vastly higher degree than would be possible if we were just sticking together third party black boxes.
 
* I would refer you to your comment about unsolved mystical issues threads on the HS forum, and point out the almost complete lack thereof on our forums. That's probably not an accident. I would also refer you to the gigantic effort that HS required to update their architecture a few years back, whereas we have made huge improvements to the product with very little disruption or complications. This is because we control what can be done, and we don't expose our internals to the customer at all. So we can change the internals as required without affecting them. So, believe me, we know what we are doing here and we didn't do all of that coding just for fun.
 
And they absolutely do not want any situation in which their getting a call at 9PM on Saturday night is because of their customer's ISP having issues. 'Thugh cloud' to them would be just one more big thing that can fail, but which isn't required to do the job. Therefore it's surplus to requirements mostly.
 
There's nothing wrong with cloud based stuff on the periphery of course. The same applies to any web based information sources such as weather or metadata, or remote connection services as mentioned above.

 

[wkearney99]

It's indeed true that with what's been offered thus far there's been little to entice the larger consumer market toward and on-going revenue stream with regard to home automation.
 
Amazon needed to build a worldwide network of servers to support their mercantile business.  Along the way they discovered if they were going to build such a thing for themselves it might also serve as a salable product to others.  Thus AWS came to be.  The tools for using it are amazingly robust and wide-reaching.  This is not just VM hosting, it's capable of offering much more.  To whit, there's the AWS Lamba service, which in effect gives you a million free transactions a month.  This is what Alexa voice skills (can) utilize.  This is a big step forward away from previous hosting models.  At the same time IBM has been busy creating a similar effort they call Bluemix.  Also, not just a bare VM but an actual code hosting environment (Wink NodeRed uses it).
 
It's a mistake to think the cloud is going to be a simple metadata resource.  It hasn't been only that for several years.
 
I point out of the AWS story because it's the kind of thing that offers a way to glue together a lot of messy bits.  Precisely the same kind of thing ill-fated efforts like IFTTT have been attempting to provide.  A robust framework on which 3rd parties can reliably build, and operate with actual contractual terms for support and uptime. 
 
Sure, once upon a time there was a lot to be said for trying to nail all the Jello to your own trees, and write all of the code.  There will always been a need for that, but it's not the kind of solution that's going to move HA forward.

 

[Dean Roddey]

I don't think that's really a very useful comparison. AWS is a code hosting service. That's it basically. You still have to write all of the code that makes up your product. But of course it now has a vastly more complicated problem of doing it remotely, when all of the stuff you need to control is somewhere else. It doesn't make it easier, it makes it harder. And when that product is very complex to begin with, the difficult compounds.
 
And if that code isn't high quality, it doesn't matter what the up time of the server is. It's the code quality that matters. I'd point you to the current situation at Smarthings. They have a cloud based setup, but the code itself is apparently not of sufficiently high quality, to the point that some of their big third party app developers are pulling their products.

 

[123]

@Dean
I'll be the first to concede the remote-access aspect of openHAB's 'cloud service' is not what comes to mind as a key benefit of the "cloud". However, to their credit, they also leverage it to provide access to their openHAB IFTTT channel.
 
I read a few threads in the Smarthings forum and it reminded me of Vera; bugginess, plus unhappy users and developers.
 
@rismoney
1600 lines of Java for a UPB driver ain't all bad! The two largest source files for Premise's UPB driver contain 3769 and 2424 lines of C++. Sure, that includes comments, blank lines, and lines containing nothing more than braces but there you have it. Add a few hundred lines in the header files and an XML file and you've got many thousands of lines of code. The driver can discover devices but I doubt that feature adds thousands of lines of code to the project.

 

[BobS0327]

I'd point you to the current situation at Smarthings. They have a cloud based setup, but the code itself is apparently not of sufficiently high quality, to the point that some of their big third party app developers are pulling their products.

 
Yea, I also saw that on the ST forum.  One of the end users in the ST community decided to remove his Rules Manager script from his GitHub repo. He felt that the ST hub wasn't stable enough to run his script.  Thus, the script is no longer available to other ST end users.  He claims it's a ST firmware upgrade issue.

 

[Dean Roddey]

I think it's more than just him. Jkmonroe on our forum said that most of the apps he uses have been withdrawn. And, from what I understand, that Rules Manager thingy filled a fairly large hole in ST's functionality. 

 

[rismoney]

Yea, I also saw that on the ST forum.  One of the end users in the ST community decided to remove his Rules Manager script from his GitHub repo. He felt that the ST hub wasn't stable enough to run his script.  Thus, the script is no longer available to other ST end users.  He claims it's a ST firmware upgrade issue.

 
If you are going to use an opensource module, always fork it.

 

[Dean Roddey]

BTW, for those folks who do like horse drawn or hand cranked automation products, we are having one of our bi-yearly 25% off sales from the 15th to the end of the month.

 

[pete_c]

Here is a new smart hub that mentions no dependencies on the cloud called Protonet ZOE.
 
It due out at the end of 2016. 
 
For now you get a Protonet ZOE and voice drop for $249 or just a Protonet ZOE for $169.
 
List price will be $399. 
 
They are currently at 200% of their 100K goal.  I am impressed with the plug n play here. 
 
https://vimeo.com/159017509
 

 
Protonet sells a private cloud server.  You can read about it here.

 

[wkearney99]

Big 3rd party and Smart Things.  Now there's a loaded bit of verbiage.  
 
Smart Things is indeed not without their complications.  It does appear they're doing a lot to address issues.  Time will tell if it helps.  What's been interesting is their embrace of open integration and support for experimentation without a lot of hassles.