Decentralized IoT Technical Design Document
- Tech Stack Requirement for : Building a API framework for communication between two IoT devices.
Language to build the APIs that enable communication between the IoT devices: Python Web Framework: Flask (v1.0.2) Source Code Control: Git (Github) IDE: PyCharm (v2018.3.1) Hardware requirements: Minimum 512 GB RAM, 1GHz single-core CPU (Based on Raspberry Pi Zero) 2. Accounts and Infrastructure 2.1 Development Dev Environment Operating System: Raspbian (Debian based operating system) Test IoT device : Raspberry Pi Zero Ws for simulation of the scenarios along with tactile push buttons and LEDs. 2.2 Production Production Environment Operating System: Raspbian (Debian based operating system) IoT device: Raspberry Pi Zero W Data Sources, Models, Timing 1.1 Data Sources Data sources will the IoT devices which keep sending the status updates to manager every time state change occurs. On every keepalive from the device, the device piggybacks the metadata to the manager 1.2 Timing For every predefined time interval (defined in the initial YAML configuration file), the device sends a keep alive message to the manager. With the keep alive, the metadata of the device is also piggy backed. For the current implementation, the manager does not report the state of all the devices to the cloud. System Architecture Diagram
Deployment Methodology Compiler which is a python script parses the IoT scenario YAML configuration and creates as many python files as many devices in the cluster. The compiler also creates a python file should be running on the manager. The name of the compiler script is Compiler.py which parses the config file named Config.yaml All the above files are bundled in a folder for each device and contains the parameters for each file. For example for parking lot, it would have Entry.txt,Exit.txt,Lot.txt While deploying the framework, the folder will have to be uploaded to the devices and running the properties in the corresponding *.py file will start the Flask servers that will help communicate with other devices. In case of the simulation of one of the scenarios, along with the Flask applications, a script that monitors GPIO will also run. The GPIO process is necessary to record input and send output for the simulation.
Note: The project does not focus on the methodology of how to deliver the folders to the devices. That is a solved problem in the domain of continuous integration/continuous deployment.