Home Automation. The holy hipster and geek grail. I have played with it. I have tried. I have failed. But today I am proud to have a solution I can truly endorse. So join me on this journey. This series will explain my solution, in excruciating detail. In the hope that I can learn from you while I am explaining. This series will be filled over time with more and more articles. But now, let’s talk about philosophy. The Why. Soon you will see the What and How. One promise, or the TL;DR: It is all 100% Open Source.
Well, almost. I have integrated some quite non-open things but always in an Open Source Way.
To wet your appetite, Here’s the simple thing you will see on a display hanging on the wall in my apartment:
What you see here is the main screen of my Dashboard. From here you can switch Light Scenes, roll up or down the projector screen, send my Roomba to clean my place and observe the current outside temperature. This all runs from a Raspberry Pi 2 with FedBerry (23 ATM), the Mosquitto MQTT Broker, a Node-RED instance with the node-red-dashboard, controlling my Philips HUE lights, a Nanode, an ESP8266 with a Dallas DS1820 temperature sensor. Too much Techno-Babble? OK. That’s all going to be explained in the other parts of this series, so relax!
Philosophy. Or the Zen of HA
Home Automation truly is a fascinating field. But it’s also filled with land mines, frustration, lack of interoperability, competing standards, non-maintained cuteness and hype. A lot of hype.
Here are the guiding principles for what I consider to be a good Home Automation (HA) solution.
- Locality Everything happens on the local network. There is no need at all for any external connections or services. This is where I protect my privacy. And guarantuee (well, shift the blame 😉 ) that I and only I am in control. Remember the NEST fiasco? Not going to happen here.
- Simple The solution must be simple to use, maintain, extend and install.
- Open The parts of the solution I interact with must be 100% open. Ideally Open Hardware and Open Source. But at least accessible using a 100% Open Source solution.
- Sustainable This means two things. It uses modules and parts that are actively maintained and simple to use for a long time. And it means that the solution doesn’t need my daily attendance. That I don’t have to be bothered with updates, changes, downtime and frustration. This is the unsolved part, to be honest.
- Rock Solid The solution should be extremely stable. And in case of problems there must be a (manual) fallback solution for all parts.
Packed with failed experiments, the lessons I’ve learned at Red Hat about long-term stability and sustainability of Open Source and the vibrant community of fellow geeks sharing their experiences and solutions, I came up with the following approach:
- Low Level Use “things” in the widest sense that can always be physically controlled. With switches, buttons or by simply pulling the plug.
- Communication All parts talk MQTT. Period. They send status updates and state changes via MQTT. They receive commands via MQTT. If they cannot do that directly, there must be an open source solution that does the translation between whatever protocol it uses and MQTT.
- Configuration Now that we only have MQTT messages to send and receive, we use Node-RED as the rules engine. Where we plug the parts together and come up with interactions as needed.
- Simplicity With a nice UI using node-red-dashboard in a browser, I can use whatever device that has a browser to control my apartment. Cheap Android tablets mounted to the wall, my smartphone or tablet. Or some voice controlled device, maybe? Mycroft?
So there you have it. The blueprint for my truly open HA solution with flexibility and agility on all layers. Sounds too good to be true? I thought the same. And in some parts it still is. But for geeks with a bit of time, I can offer quite a working solution. Stay tuned! On to the next parts, where I will dive into the how and what.
Comments? Always welcome!
Or right here (but I’m more responsive on G+ and Twitter).