Intelligent Workloads at the Edge: Deliver cyber-physical outcomes with data and machine learning using AWS IoT Greengrass

✍ Scribed by Indraneel Mitra, Ryan Burke

• Publisher: Packt Publishing
• Year: 2022
• Tongue: English
• Leaves: 374
• Category: Library

✦ Synopsis

Explore IoT, data analytics, and machine learning to solve cyber-physical problems using the latest capabilities of managed services such as AWS IoT Greengrass and Amazon SageMaker

Key Features

• Accelerate your next edge-focused product development with the power of AWS IoT Greengrass
• Develop proficiency in architecting resilient solutions for the edge with proven best practices
• Harness the power of analytics and machine learning for solving cyber-physical problems

Book Description

The Internet of Things (IoT) has transformed how people think about and interact with the world. The ubiquitous deployment of sensors around us makes it possible to study the world at any level of accuracy and enable data-driven decision-making anywhere. Data analytics and machine learning (ML) powered by elastic cloud computing have accelerated our ability to understand and analyze the huge amount of data generated by IoT. Now, edge computing has brought information technologies closer to the data source to lower latency and reduce costs.

This book will teach you how to combine the technologies of edge computing, data analytics, and ML to deliver next-generation cyber-physical outcomes. You'll begin by discovering how to create software applications that run on edge devices with AWS IoT Greengrass. As you advance, you'll learn how to process and stream IoT data from the edge to the cloud and use it to train ML models using Amazon SageMaker. The book also shows you how to train these models and run them at the edge for optimized performance, cost savings, and data compliance.

By the end of this IoT book, you'll be able to scope your own IoT workloads, bring the power of ML to the edge, and operate those workloads in a production setting.

What you will learn

• Build an end-to-end IoT solution from the edge to the cloud
• Design and deploy multi-faceted intelligent solutions on the edge
• Process data at the edge through analytics and ML
• Package and optimize models for the edge using Amazon SageMaker
• Implement MLOps and DevOps for operating an edge-based solution
• Onboard and manage fleets of edge devices at scale
• Review edge-based workloads against industry best practices

Who this book is for

This book is for IoT architects and software engineers responsible for delivering analytical and machine learning–backed software solutions to the edge. AWS customers who want to learn and build IoT solutions will find this book useful. Intermediate-level experience with running Python software on Linux is required to make the most of this book.

Table of Contents

1. Introduction to the Data-Driven Edge with Machine Learning
2. Foundations of Edge Workloads
3. Building the Edge
4. Extending the Cloud to the Edge
5. Ingesting and Streaming Data from the Edge
6. Processing and Consuming Data on the Cloud
7. Machine Learning Workloads at the Edge
8. DevOps and MLOps for the Edge
9. Fleet Management at Scale
10. Reviewing the Solution with AWS Well-Architected Framework

✦ Table of Contents

Cover
Title page
Copyright and Credits
Contributors
About reviewers
Table of Contents
Preface
Section 1: Introduction and Prerequisites
Chapter 1: Introduction to the Data-Driven Edge with Machine Learning
Living on the edge
Common concepts for edge solutions
Bringing ML to the edge
Tools to get the job done
Edge runtime
ML
Communicating with the edge
Demand for smart home and industrial IoT
Smart home use cases
Industrial use cases
Setting the scene: A modern smart home solution
Hands-on prerequisites
System 1: The edge device
System 2: Command and control (C2)
Summary
Knowledge check
References
Section 2: Building Blocks
Chapter 2: Foundations of Edge Workloads
Technical requirements
The anatomy of an edge ML solution
Designing code for business logic
Physical interfaces
Network interfaces
IoT Greengrass for the win
Reviewing IoT Greengrass architecture
Checking compatibility with IoT Device Tester
Booting the Raspberry Pi
Configuring the AWS account and permissions
Configuring IDT
Installing IoT Greengrass
Reviewing what has been created so far
Creating your first edge component
Reviewing an existing component
Writing your first component
Summary
Knowledge check
References
Chapter 3: Building the Edge
Technical requirements
Exploring the topology of the edge
Reviewing common standards and protocols
IoT Greengrass in the OSI model
IoT Greengrass in ANSI/ISA-95
Application layer protocols
Message format protocols
Security at the edge
End devices to your gateway
The gateway device
Edge components
Connecting your first device – sensing at the edge
Installing the sensor component
Reviewing the sensor component
Connecting your second device – actuating at the edge
Installing the component
Reviewing the actuator component
Summary
Knowledge check
References
Chapter 4: Extending the Cloud to the Edge
Technical requirements
Creating and deploying remotely
Loading resources from the cloud
Packaging your components for remote deployment
Storing logs in the cloud
Merging component configuration
Synchronizing the state between the edge and the cloud
Introduction to device shadows
Steps to deploy components for state synchronization
Extending the managed components
Deploying your first ML model
Reviewing the ML use case
Steps to deploy the ML workload
Summary
Knowledge check
References
Chapter 5: Ingesting and Streaming Data from the Edge
Technical requirements
Defining data models for IoT workloads
What is data management?
What is data modeling?
How do you design data models for IoT?
Selecting between ACID or BASE for IoT workloads
Conceptual modeling of the connected HBS hub
The logical modeling of the connected HBS hub
The physical modeling of the connected HBS hub
Designing data patterns on the edge
Data storage
Data integration concepts
Data flow patterns
Data flow anti-patterns for the edge
A hands-on approach with the lab
Building cloud resources
Building edge components
Validating the data streamed from the edge to the cloud
Additional topics for reference
Time series databases
Unstructured data
Summary
Knowledge check
References
Chapter 6: Processing and Consuming Data on the Cloud
Technical requirements
Defining big data for IoT workloads
What is big data processing?
What is domain-driven design?
What are the principles to design data workflows using DDD?
Designing data patterns on the cloud
Data storage
Data integration patterns
Data flow patterns
Data flow anti-patterns for the cloud
A hands-on approach with the lab
Building cloud resources
Querying the ODS
Building the analytics workflow
Summary
Knowledge check
References
Chapter 7: Machine Learning Workloads at the Edge
Technical requirements
Defining ML for IoT workloads
What is the history of ML?
What are the different types of ML systems?
Taxonomy of ML with IoT workloads
Why is ML accessible at the edge today?
Designing an ML workflow in the cloud
Business understanding and problem framing
Data collection or integration
Data preparation
Data visualization and analytics
Feature engineering (FE)
Model training
Model evaluation and deployment
ML design principles
ML anti-patterns for IoT workloads
Hands-on with ML architecture
Building the ML workflow
Deploying the model from cloud to the edge
Performing ML inferencing on the edge and validating results
Summary
Knowledge check
References
Section 3: Scaling It Up
Chapter 8: DevOps and MLOps for the Edge
Technical requirements
Defining DevOps for IoT workloads
Fundamentals of DevOps
Relevance of DevOps for IoT and the edge
DevOps challenges with IoT workloads
Understanding the DevOps toolchain for the edge
AWS Lambda at the edge
Containers for the edge
Additional toolsets for Greengrass deployments
MLOps at the edge
Relevance of MLOps for IoT and the edge
MLOps challenges for the edge
Understanding the MLOps toolchain for the edge
Hands-on with the DevOps architecture
Deploying the container from the cloud to the edge
Summary
Knowledge check
References
Chapter 9: Fleet Management at Scale
Technical requirements
Onboarding a fleet of devices globally
Registering a certificate authority
Deciding the provisioning approach
Managing your device fleet at scale
Monitor
Maintenance
Diagnose
Getting hands-on with Fleet Hub architecture
Building the cloud resources
Deploying the components from the cloud to the edge
Visualizing the results
Summary
Knowledge check
References
Section 4: Bring It All Together
Chapter 10: Reviewing the Solution with AWS Well-Architected Framework
Summarizing the key lessons
Defining edge ML solutions
Using IoT Greengrass
Modeling data and ML workloads
Operating a production solution
Describing the AWS Well-Architected Framework
Reviewing the solution
Reflecting upon the solution
Applying the framework
Diving deeper into AWS services
AWS IoT Greengrass
AWS IoT services
Machine learning services
Ideas for further proficiency
Summary
References
Appendix 1 – Answer Key
Chapter 1
Chapter 2
Chapter 3
Chapter 4
Chapter 5
Chapter 6
Chapter 7
Chapter 8
Chapter 9
Index
Other Books You May Enjoy