@Ben, having worked at a software company in may storied past, I completely agree with the philosophy and it makes sense. However, I would hope that on the roadmap (sooner rather than later) is some very basic functionality that can be handled locally…and I’m OK knowing that it may require a new Hub.
But, as you read through the messages here in the community, there are some critical tasks being handled by the Hub, which I think is a testament to the faith everyone has in ST. But, as an IT Director who is used to building disaster recovery scenarios that include stages with “minimal functionality to get us going”, I think ST should see if there are any simple, critical, or reduced functionality (read: simple) tasks that can be offloaded to the hub in an emergency situation. And, to be honest, that situation could be a ST cloud outage in addition to an internet outage. (Seriously, who knows what the hell Comcast and all these telcos have up their sleeves once they bulldoze through their version of Net Neutrality!)
I’m very much looking forward to getting up and running with my new Hub and devices this weekend, and I’m looking at the long term on this. I think that is why you are hearing the feedback like this from the Community.
Thanks @brianlees - This is from the same page in the documentation:
That said, we are actively considering implementation scenarios whereby we can distribute SmartApps to and execute SmartApps locally on the hub.
In all cases, we obviously recognize the critical scenarios where a loss of communications with the SmartThings Cloud could have a degrading impact on critical, local use cases and are being deeply thoughtful on how we minimize the risk of disruption.
It will require a new hub, as you mention. There is no solid timeline on this but we have prototypes working.
Some logic must be brought home. When using ST as a home security appliance (along with awesome convenience features), the Hub should know what state it is/should be in (Home, Away, Night, etc) and if something not allowed happens (door opens, window opens, motion, etc). It should be able to stand alone and signal a siren, lights, etc… A temporary Internet connection outage should not make you less safe when using ST. Maybe you lose some non-essential functionality when your Internet connection is lost, but there exists some consistent basic functionality that should be processed in the Hub during such an event.
While I completely understand the reasoning behind the brains being in the cloud it is a bit disconcerting. I’m a very new user and have no programming experience but was able to use others contributions to build support for an “unsupported” device. Without the environment set up by ST I’m not sure that would have been available to me. That said I did start looking at other home automation software and less expensive hub devices as there is more functionality available (home theater control) and the controls reside locally. Having to rely 100% on the cloud is a problem and precludes using this as a security system or, to some extent, for HVAC controls.
The need for new hardware rings a bit hollow to me as it would seem that having a local “mirror” on the home network would allow the hub to function locally in the case of an internet outage. Or whenever the user is on the local network.
Maybe the Galaxy S6 will have ST functionality to control the hub built in,
Hello all.
I love the concept smartthings and am currently developing a mvp product at work using it.
However I find the concept of relying on the cloud or even the internet at little alarming.
My city was hit in 2010 and again in 2011 by a significant earthquakes which left compromised infrastructure and services. things are pretty much back to normal now but cooking dinner every night on a bbq for the family, power cuts and water outages gave me a new appreciation for redundancy, fail safe and reliability issues in my work and personal life.
Hence it would be nice if the hub and my smart devices would still work with no internet
One thing you need to keep in mind when you invest in home automation is that home automation is actually an industry that started with business. Since this is DIY you need to think of your home as a business. If you fear the power going out, install an UPS and use battery devices for security related things. If you fear your internet going down, then buy a router that can do backup 4G and get a monthly device service plan from your carrier (like $10, uses your pooled data connection).
SmartThings can’t solve all your problems but the best thing about it is that it’s programmable. So where other devices you get a cookie cutter feel, in ST if you’re truely a DIY’er then you’ll really succeed. It’s not for everyone.
The problem is that the emphasis is on automation. That like saying that steak is all about the sizzle. I first need to be able to do something as simple as turning on and off my lights reliably. If I can’t do that then I can’t use SmartThings for more than a demo and have to look to alternatives.
why would we need a new hub? If my phone/tablet is connected to the same local network as my hub then why not allow the Smarthings APP handle local processing?
In my humble opinion, the Internet of Things means I can get data/status from and control a thing (washer,dryer,fridge) via the internet. If there is no internet, than its just a dumb fridge keeping my food cool manually/automatically. It still work works, its just not as cool/convenient.
Everything I have installed has a manual mode (except for sensors). So,I can still turn on the lights, heat my house and wash my clothes. Even when the internet is out. I can’t do it from a million miles away though
We’re quickly getting to where the Internet is as reliable as a Utility. Much like the electricity that’s hooked to each of our houses. We are also getting to where each server/app/company that is on the other end from us is as reliable as a Utility. It just takes time to build that infrastructure.
It wasn’t that long ago I would jump for joy when I got my 2400 baud modem to finally connect to some server, somewhere, so I could get on a BBS. Back then we’d all bitch about how often you would get kicked off by a moderator. Yes, a human could decide if you were good enough to use the Internet.
Anyway, I digress…
All that being said, I too do want local control, but I created my automated home to be totally functional without the Internet, or Hub for that matter. First rule of automation, is to first make it work manually. First rule of agility in automation, is to maintain the ability to do the automated manually.
You’re right. Internet connectivity is just an extra convenience. A smart washer allows me to set up schedule and wash cycles and it will execute them without Internet just fine. That is not the case with SmartThings, where none of my schedules will execute without Internet connectivity.
The point is, it’s ok to forgo remote control and notification without Internet, but losing even basic smart home functionality like scheduled events is not what an average consumer would accept.
I dove into this HA lark in the middle of the bad times a month or two ago when nothing worked for several days. I very quickly modified my stance to make sure that anything I make smart also works dumb. I have physical Z-Wave switches/dimmers for all of my lights (Smart bulbs are usually OK in that if you toggle the light switch they will default to full on, but you lose dimmer capabilities), I still have manual control of my garage, when I get thermostats they will work as dumb thermostats. My goal is that if the internet went away forever tomorrow my home would work just as well as it did before SmartThings, even if i do have to go through the horrific inconvenience of turning 5 lights on downstaits at night manually
I also agree that the Internet is evolving into a Utility with that level of reliability. I can’t actually remember the last time my Internet went down. I think where we are now is similar to the early days of electricity where we would keep a good supply of candles on hand to get through the frequent power cuts. In 10 years time we won’t need our “candles” any more.
However, I do still want to see some local processing, purely form a latency and reliability standpoint. My single hub running my few rules is a lot less likely to get bogged down than a cloud infrastructure servicing thousands of homes. It still blows my mind that I can turn on all my downstairs lights by:
Pressing an icon on my iPad
That communicates with a Heroku server somewhere in some cloud probably in California
That then communicates with the SmartThings cloud somewhere else
That communicates with my hub over the Internet
That finally uses Z-Wave to signal my lights to come on
If I sat and thought about it for a while I would be hard pressed to come up with a more contrived replacement for a simple light switch!
5 Likes
tgauchat
(ActionTiles.com co-founder Terry @ActionTiles; GitHub: @cosmicpuppy)
35
How many entrepreneurs, programmers, smart phones, internet connections, … does it take to turn on a light bulb?
This notion that there a far off Internet we “access” and that we must depend on some sort of far off cloud to turn on and off lights shows a fundamental lack of understanding of the very nature of connectivity which should depend only on the two end points in the relationships.
if all you want is trophy home automation then there is absolutely nothing wrong with using IFTTT and dependent on a server in Tierra del Fuego. And that’s OK.
But if you want real infrastructure and want to use 20th century signaling rather than 19th on/off electric wires then you can’t require every entity along the path to work perfectly. Messages must be sent directly from the “switch” to the “bulb” or ,at worst, through a local hub.
Absolutely! The idea of sending a packet on a journey across the entire continent and back to turn on a light bulb few feet away is simply preposterous, no matter how many people pride themselves of this “achievement”.
2 Likes
tgauchat
(ActionTiles.com co-founder Terry @ActionTiles; GitHub: @cosmicpuppy)
40
Your observation is fundamentally “common sense” (and I share the sentiment), but the world has become a complex place. Even adding one layer of complexity (of certain sorts…) suddenly creates a legitimate need for “cloud” processing.
The example I have in mind is natural speech recognition (e.g., The Ubi, Ivee, Cubic, Amazon Echo, Google Now, Siri, Cortana …). What is the cost to put the processing power necessary for this particular functionality in a local hub? What is the net benefit?
About 20 years ago, Oracle “invented” the “Network Computer” (i.e., a dumb terminal for a virtual desktop session running on a remote server) http://en.wikipedia.org/wiki/Network_Computer – it did not sell well, though a variation was widely used within Sun Microsystems. More recently, the Chromebook was released which is pretty much the exact same thing – I don’t think they are very popular either, I guess.
My point is that there will always be a set of architects that believe in minimizing computing resources at the client (user) endpoint, just due to net efficiency. Perhaps it is mathematically true that the network cost is trivial in this equation; but the cost of network outages and security risks is harder to quantify.
