The system based on micro services architecture, and it is:

• event driven design
• platform agnostic

flexibility and resilience

EdgeWrapper ==> MicroService:
mode:
    - solo mode (single device no distributed system)
    - network mode (multiple devices, as distributed system)
context: Local Network
Components:
    - software services tree
    - Web server
    - API (end points for every PyService)
    - Dashboard
    - Admin (for every PyService)

EdgeService ==> SoftwareService:
context: Host
Components:
    - API registry to EdgeWrapper (using Flask & Blueprint)

Services manager, provide API, B.O, WebUI, admin, dashboard.

Sensor data is collected by a Device Service from a thing, and published in Mqtt-service.

Device-service provide API

The MQTT protocol provides a lightweight method of carrying out messaging using a publish/subscribe

Configure device machine type by env var:

EDGE_MACHINE_TYPE=pc_linux

Docker base image by machine types

balenalib/raspberrypi3-python:3

balenalib/raspberrypi3-debian:jessie

balenalib/raspberrypi3-debian:stretch-run-20190612

x86:

• pc_linux
• pc_win
• mac

ARM:

• rpi

La clé se n'est pas le code, mais l'architecture

CustomProject
    /-- PyEdgeIoTFramework
    /-- CustomService
    docker-compose.yml

edge-server
    folder: PyEdgeIoTFramework
    env_avr: server
device-service
    folder: PyEdgeIoTFramework
    env_avr: device
mqtt-service
    image: mosquitto

custom-service
    folder: CustomService

Commands:

• ls: pour lister les fichiers du répertoire.
• cd: pour changer de répertoire.
• mkdir: pour créer un répertoire.
• grep: pour rechercher un mot.
• sort: pour trier des mots.

• ToDo: Add PySse service (like mqtt)
• ToDo: Add PyMplayer service (like omxplayer)
• ToDo: manage all exception and make a loop to retry action failed