EcoHub

A simulation of an internet of things (IoT) system with smart devices and a central controller communicating with them as their state changes.

This project aims to incorporate concepts from OOP Python, concurrent programming with the asyncio module as well as with threads, and functional programming concepts.

Demo

How it works

• Smart devices (smart bulbs, smart thermostats, smart cameras) connect to a controller and, as their state changes over time (temperature changes, motion detection, humidity changes), they send updates to the controller in the form of packets containing payloads.
• The controller receives incoming packets via a asyncio.Queue. The controller is on standby until it receives a packet, without blocking the devices from sending packets. Each packet from a device is processed sequentially by the controller.
• Using a module called analytics engine, packets are turned into payloads that the controller stores for each device connected and updates the device state accordingly. For example, when a smart thermostat connects to the controller, it sends a payload. The controller receives the payloads and stores a reference of the object representing the thermostat device, and stores the received parsed payload. When the device sends a packet again, this stored payload is updated.
• The analytics engine then filters out received payloads to determine critical events, i.e. if any action needs to be taken by the controller. For example, if a smart camera detects motion, the analytics engine will flag this event as critical and the controller will take action by sending a command to the camera to take a snapshot. Critical events are then sent to the devices through a method on the device object. When the device executes the command, it prints a corresponding message to the console.
• The analytics engine also computes metrics for every received payload: average house temperature, average humidity level, total number of devices connected. These metrics are printed to the console.
• For every received parsed payload, the controller calls a storage worker, running in a separate thread, to log the payload to a file.
• The storage worker has a queue (queue.Queue) that the controller puts payloads into for logging. The storage worker thread blocks on this queue until a payload is available to log, then logs the payload to a file (every payload is logged sequentially), flushing the file (saving the file and writing to disk) after every write.
• When the program is terminated (via CTRL+C), the controller inserts a None value into the storage worker's queue, signalling it to terminate.

Diagram

Below is a high-level diagram of the EcoHub system architecture. Note that the notation used does not conform to any standard (e.g. UML) and is for illustrative purposes only 😅.

Concepts learned

See the wiki page for explanations of concepts learned and architectural choices made.