Most IoT projects today lean heavily on Python, C, or Node.js, and that’s fine. But during my recent academic paper selection process, I came across “The Benefits of Tierless Elixir/Potato for Engineering IoT Systems”, and it completely shifted how I think about building IoT architectures.
The paper raised a question that stuck with me:
why do we keep separating the logic, runtime and UI layers in IoT systems if functional, tierless architectures can unify everything?
That curiosity, combined with the influence of a professor(Adolfo Neto) who strongly advocates for functional programming and the BEAM, pushed me to test this idea in practice.
So I built a complete IoT prototype using Elixir, Circuits, Raspberry Pi, and Phoenix LiveView. And honestly? It felt like IoT the way it should be: supervised, fault-tolerant, reactive, and consistent from the edge to the dashboard.
The Prototype: What I Actually Built
My goal was simple:
Build an IoT system end-to-end using only Elixir, from the hardware to the backend.
That resulted in a two-part monorepo:
- A Raspberry Pi running Elixir (no Nerves this time, just Elixir Circuits)
- A Phoenix LiveView dashboard receiving sensor data in real time
The device communicates via Phoenix Channels/WebSockets, sending temperature, light level, and button events, while receiving commands to control a buzzer.
Architecture Overview
If a sensor process fails: only that process restarts.
If the network blips: the BEAM keeps the system stable.
If the dashboard disconnects: LiveView reconnects automatically.
This is where the “future feeling” starts to show.
Circuit Setup (Everything on the Pi4)
Hardware used:
- DS18B20 (1-Wire temperature sensor)
- LDR + Capacitor (for RC timing)
- Button (pull-up GPIO)
- Active buzzer + transistor
The LDR timing loop runs in a separate process so it doesn’t block WebSocket communication, a natural fit for the actor model.
Why Elixir Feels Like the Future of IoT
After building this, a few conclusions became obvious:
- The BEAM VM solves problems IoT devs fight daily
- Concurrency is natural, not bolted-on
- Fault tolerance is built-in
- Real-time dashboards require zero JS
- A single language from hardware to UI reduces mental load
- Tierless architectures aren’t just academic, they’re practical
This cohesiveness made the whole experience feel weirdly futuristic.
Not flashy.
Not hyped.
Just… right.
What’s Next?
- Implementing a Nerves firmware version
- Adding more sensors
- Testing distributed BEAM cluster of Pis
- Applying Potato tierless concepts end-to-end
- Publishing a deeper academic report
Elixir isn't mainstream in IoT, but maybe it should be.
This small prototype convinced me that a unified, functional, message-driven approach has the potential to dramatically simplify embedded systems.
If you're curious about IoT, the BEAM, or tierless architectures, give Elixir a try.
It might surprise you the same way it surprised me.
Try It Yourself
tuliocll
/
elixir-iot-sample
A complete IoT example using Elixir, Phoenix LiveView, and Raspberry Pi.
Elixir IoT Example 🚀
Elixir IoT Example is a monorepo demonstrating how to connect a Raspberry Pi to a real-time Phoenix LiveView dashboard using Phoenix Channels, WebSockets, and Elixir Circuits.
It includes:
- 🌡️ Real-time temperature readings (DS18B20)
- 💡 Light level measurement using RC timing (LDR)
- 🔘 Physical button input
- 🔔 Remote buzzer control through WS commands
- 📊 Web dashboard with auto-updating LiveView
This project serves as a practical, didactic example of building IoT systems entirely in Elixir.
Requirements
- Elixir 1.17+
- Erlang/OTP 26+
- Node.js 16+ (for assets)
- Raspberry Pi 4 / 3 (tested with Pi4)
- Erlang + Elixir installed in the Rasp
- Enable 1-Wire in Rasp (See doc/1-wire.md)
🏗️ Architecture and Circuit
Running server
cd server
mix deps.get
mix phx.server
Open:
http://localhost:4000/dashboard
📡 Raspberry Pi (Elixir Circuits)
Running
# Clone the repo on the rasp and:
cd rasp
mix deps.get
iex -S mix
The device will:
- …
Top comments (0)