Openhab is the Intranet of things, therefore why there is no specific platform for running it, simplier and easier for the masses.
A specific SD Card image will all with several configurations USB stick or Razberry or all in one with HAbmin and with debuggers, mysql and +++ pre-configured, or a specific hardware being managed directly by openhab leads?
Internet of Things = Vera lite - Smarthings - all the rest…
Intranet of Things = Openhab (if you want IOT you also can).
Why we need to share our data with internet (maybe we want to have some privacy)… And what happens with the Internet of Things in case there is no internet… Intranet of things will always work. But the InteOT will not.
Whilst I am a little confused about your “intranet of things” argument.
I actually like the idea of an image that has a standard working openhab install, that is configured to autostart, and optional components all easily enabled.
An alternative is a maintained script that will download and configure openhab (ie permissions, autostart, etc)
RetroPie, which is a project to use a Raspberry Pi as a gaming emulator, does this quite well, they have a script to build it as well as a prebuilt SD image, that can be loaded on to an SD and booted on a RPI http://blog.petrockblock.com/retropie/#retroproject
I also like the idea of a preconfigured SD card, but I would also like to keep openHAB independent of any specific hardware and support the different platforms equally well. That’s why putting a lot of effort specifically in a RaspPi image is not on OUR list, but it would of course be very welcomed if being done by the community. The scripts that @peter mention sound like a good starting point.
In general, I am rather looking at things like Docker container or Snappy apps; if we had such images automatically built, this could be used on many platforms; so we could have an RaspPi SD card image, which simply includes the Docker container in it. This might be the best way forward.
Actually, I am working on a docker image of Openhab 2.
At the moment, I have this huge shell script that installs the server part specifically for my Raspisberry Pi setup, but after gaining a bit more experience with docker, it will convert it to a Dockerfile.
I absolutely like the subject of this thread. Even if it is one of the major topics right now… The interpretation of ”Internet of Things” of the most vendors is something it totally dislike.
Why should my washing machine or my soundbar talk directly to some kind of (proprietary in the most cases) cloud service which is than optional connected to some kind of cloud rule management like IFTTT ?
For me openHAB is a very flexible solution to build my vision of a ”Intranet of Things" implementation, where I do control which parts are visible via the cloud… If any…
I agree here with the “intranet” of things. So far on my install I have tried to limit how much requires internet access. Granted we have FIOS here in FL so I have not actually had an outage in years. But it can still happen, and I can tell you now, the WAF (Wife Acceptance Factor) of the home automation system would go to 0 if she can’t turn on a light because the internets down. I would have to put all my fancy toys back in my room.
For me it is not just the risk of an outage. It is the complete set of possibilities, so for instance the risk of misusage of the provided data and the risk to get hacked. There comes a risk with every cloud service I use. One for washing machine, one for the vacuum cleaner, another one for the heating etc…
I personally prefer to limit the number of those services, but a lot of new products that have been introduced to the market the last months follow a different philosophy. A philosophy that’s no mine
agreed! i also like to keep my systems simple. Only supply the data that is NEEDED. so my sensors for washing machine and what not ONLY report simple data (on/off) and only to a local point. No need having someone turn my dryer on for hours on end while im away and burning it out. or trying to flood my house lol
please also have a look at https://hub.docker.com since there are already some ready to use images. Don’t know about their state though. It would be great to have some kind of docker image hierarchy (Base-Image + Java + openHAB) or alike.
Unfortunately we would have to solve the Oracle JDK agreement issue (the enduser has to confirm the license agreement manually) somehow. We would also like to take contributions in that area and feed the openHAB docker user with some nice recipes
I’d like to support this idea too. Whilst hardware independence is a great goal, don’t ignore simplicity and cost. The Raspberry Pi is cheap, fairly well connected and there is good support for addon goodies like wifi and an LCD touch screen. This makes a great cheap starter pack for learning. It isn’t what I’d choose for a production environment but heck I run asterisk on one so my phones don’t work if it fails (it hasn’t). If ‘openhab’ is designed for hardware independence I should be able to take a working configuration developed on a Raspberry PI and load it onto my unRaid server running ‘openhab’ in a docker. Not only does this make development independent of hardware it also protects the important stuff from development glitches!
This is also a great model for making a small continuous revenue stream which is also necessary.
@Topic starter:
Nothing against your privacy concerns (on the contrary), but in a extranet approach, where intranets communicate via (secured tunnels on) the internet, you are in control of your data, and the OpenHAB groups extend beyond a single intranet/location. In that regard (technically speaking) the Internet of Things isn’t such a bad name.
</nitpicking>
I agree: Privacy is important to most users, and naming the kid “intranet” might make people think the powerful concept is limited in range. I would rather go for “network(s) of things” or “networked things”. OpenHAB is a gateway between various transport layers, the IP networks only being a part of the possibilities. But “Internet” at least is a better pick than “cloud”.
And what about people who want to automate mobile plant in outback Australia where the is no internet access. This is the only option I have found that doesn’t assume internet access is available and could have many applications not necessarily “home” based. I am looking forward to seeing if it is workable. Can anyone tell me (or point me in the right direction) as to how openHAB works “offline”? WIFI with static IP addresses?
@Matt, Home automation and plant or industrial automation are two very different fields of study.
Both can use ethernet communications both wired and wireless as one way of communicating from machine to machine.
Plant or factory automation place greater value on low latency or delay of communications than home automation. Extensive error checking of the received data is also a normal requirement in industrial automation. A few TCP/IP transport industrial solutions I can think of are Profibus, DeviceNet and ControlNet. I am sure there are more. Can bus is a mature communications hardware enabled communications protocol that has chip level hardware support from most major embedded controller IC manufacturers.
For a robust publish/subscribe eco-system over TCP/IP on a local network, ROS (Robot Operating System) may be worth a look. It is up to you to decide the speed and communications reliability required for your application.
What do you mean by “offline”? In most situations openHAB is hosted on a local area network (LAN) connected to the “internet” wide area network (WAN) through a network router (often called a “gateway”). That said, openHAB can work effectively on a LAN with no gateway to the internet. The LAN could be wired or wireless or some combination of the two. For home installations, LAN IP addresses are typically assigned by a host on the LAN using a DHCP server (typically the same host as the router/gateway) so static IP addresses are not required but may be useful for some purposes.
Thanks for the replies guys. Ridge, low latency and error checking are not big issues for this project though I shall look into those you mentioned. All I require is to be able to wirelessly (less than 20m) monitor a number of sensors, control basic switching as well as switch between two live video feeds. OpenHAB appealed to me because of its simple language, offline ability and readymade, technology agnostic GUI, all of which are essential for this project. Steve I did have LAN in mind and it sounds like the way to go. Can you see it working with the above mentioned parameters?
Intranet of things will always work. But the InteOT will not.
