I have been looking into this, what precisely is your application?
While I can see Nigels point to an extent, there are big differences between the pi and a pc. For a start the pi is able to operate as both system and application side, it has its own I/O for a start. Your application is paramount to your decision. If you want to switch a light on via your phone or something, then yes go with a simple app board like Nigel suggests.
But true IOT is alot more than that, it isnt the often assumed ability to do something like turn on a light over the internet. it is supposed to be a system of control where all application side devices can be controlled via a central or in some cases, a node type system. Boards that take wifi or bluetooth and can connect to a phone or pc are fine. But they have many limitations.
A RP is not a pc, its a SOC, as such it has the ability to talk to many devices natively, it comes with most common protocols like RS232,SPI,I2C etc etc, on top of that you can plug any other other device into the USB hub if needed. The power this gives you is immense.
As I still dont know enough about the boards Nigel is talking about, i will use a Arduino as a kind of similar example. yes you can plug this hat or that hat onto it, and yes communicate with it via the web, But its a mess. Ultimately your bodging bits together. With a Pi you can communicate with an arduino or most other boards easily, dosnt matter what protocol they use.
Lets take the following example.
Over the years you have built a pic controlled garage door, a arduino Based outside light switch, and a few other bits. How would you then tie everything including your new heating system together? Sure you could try and use a pc, but your using alot of power when its not needed, you also have to have a board on the pc for this or that.
Now with a pi, not only can you talk to any of the devices using whatever protocol they use (maybe via the I/O pins or USB), but you have the ability to build a website/app on the pi to control everything. Lets say the garage has an alarm, the pi would sense this and turn the light on, dosnt matter if the alarm is pic/arduino based and the light something else. The pi can talk to most things.
While not a pc it has many of the capabilities of a pc. although depending on device, you dont always need the pi powered for the device to work. they are also low powered and small, its simple to communicate with a pi, over the internet or wifi or bluetooth, even RS232. The beauty being the system can grow as you do, with the more simple boards you get one job done and thats it, if you have several do you log into each? or can you log into a single one and control the others?
I know your in a rush to decide, but hold off. You asked what seems a simple question, but it isnt that simple. I used to get told on here alot........buy once, buy right. So the more we know about exactly how you see this project going, the better the advice we can give.
I intend to do a full write up on IOT as it has applied to me, unfortunately I dont have the kind of free time i once did. I dont want to give you bad advice, but equally I dont want you to be limited for the future.
Nigel you talk about pc's and servers for these systems, while I cant tell you where or why, I can tell you that there is a building control system currently being trialed. Whats extraordinary is, the building is actually 3 different buildings with at least several miles between them. Led light systems that turn on and off depending if someone is in the room, the light itself alters its brightness depending on ambient light. The heating system and temp control for each room of these buildings, and many other things are all controlled by a single central system.
The central system is a RPI3!!! Some of the other things like the light control units are Bluetooth meshed room to room then SPI. The heating and movement system is Sil labs/energy micro M3 Cortex 32bit based. Everything links back to a total of 4 Rpi3's. 2 are for redundancy, these are large non domestic buildings. The system is meshed and been fully operational for 4 months now.
Yes teething trouble for 3 months, but the last month or so its starting to settle. I didnt have much to do with any of the design or installation, but because of the sil labs boards I do attend and fix (normally a reset lol) problems with some of the system. The main issue at the moment is the mesh side not the IOT side. I dont begin to pretend I understand how the maths work for the routing or signals (meshing), but i honestly think projects like the Op are better served by more capable systems.
Even the Arduino is now becoming little more than a slave device, no more having to try and write a webpage on a pic!! Think how much power even a pc uses compared to a RPI, add to that the ultra low energy of the Cortex boards and the whol;e system hardly uses any power.The real power comes from the tasks being distributed rather than a single device doing all the lifting. But that is the whole Point of IOT.
I do promise to write up about true IOT, from where I stand and what I have been involved in/seen IOT seems to be misunderstood, its way way past turning a light on with a phone. The PI isnt perfect, better system exist, but they cost alot more. The current building control system is due to run 18 months, if successful it will be popping up in some pretty strange places.
I think the ultimate point is this, the PI can do anything the other board can do and more, the other board cant do a fraction of what you can do with a PI. So surely the main question then becomes...
WHAT does the OP actually want to do now and in 12 months time?
BTW OP
I am not arguing against the other board, my point is unless you talk about what you see now and in the future for your project, then its guess work from us.
EDIT
In the BCS i talk about above, every light panel in every building is connected to each other. Infact everything is connected to everything else, signal routes are not fixed, they are chosen at the point of use each time. So a heating control command may well pass through the light controller if the system sees this as the most economic/reliable/best route at that point in time. To me thats true IOT, you can alot more with a whole lot less. no bottle necks and ques.
