Skip to content

Edge Computing - Notes

Table of Contents (ToC)

Introduction

Edge computing brings computation and data storage closer to the location where it is needed to improve performance and efficiency.

What's Edge Computing?

  • Distributed computing paradigm
  • Moves data processing closer to data sources
  • Reduces latency and bandwidth use

Key Concepts and Terminology

  • Edge Node: Device that performs data processing at the edge
  • Latency: Time delay in communication
  • Bandwidth: Data transfer rate

Applications

  • IoT (Internet of Things)
  • Smart Cities
  • Autonomous Vehicles

Fundamentals

Edge Computing Architecture Pipeline

  • Data Source: Sensors, IoT devices
  • Edge Nodes: Local servers, gateways
  • Data Processing: On-site computation
  • Cloud Integration: Synchronization with cloud resources

How Edge Computing Works?

  • Data is collected from local devices.
  • Processed locally on edge nodes.
  • Aggregated or synchronized with the cloud as needed.

Some Hands-on Examples

  • Processing video feeds from surveillance cameras locally
  • Real-time analytics on sensor data in manufacturing

Tools & Frameworks

  • AWS Greengrass: Extends AWS capabilities to edge devices
  • Microsoft Azure IoT Edge: Runs cloud workloads on edge devices
  • Google Cloud IoT Edge: Facilitates AI at the edge
  • K3s: Lightweight Kubernetes distribution for edge deployments
  • EdgeX Foundry: Open-source framework for building interoperable edge computing systems.
  • TensorFlow Lite for Edge: Lightweight version of TensorFlow optimized for edge devices.

Hello World!

Hello World I: Infering ML model using Tensoflow 

# Sample code using TensorFlow Lite for edge computing on IoT devices
import tensorflow as tf

# Load the TensorFlow Lite model for edge deployment
interpreter = tf.lite.Interpreter(model_path="model.tflite")
interpreter.allocate_tensors()
# Perform inference on edge device data
# (Additional code for data preprocessing and inference goes here)

Hello World II: Task Management

# Simple edge computing example using a local Python script
def edge_computing_task(data):
    processed_data = data * 2  # Simulate data processing
    return processed_data

data = 10
result = edge_computing_task(data)
print("Processed Data:", result)

References

Lectures & Online Courses: - Edge Computing Fundamentals - Coursera by Learn Quest - Security at the Edge Specialization - Secure your Edge and IoT Devices

Research & Survey:

Books: