Edge Computing Patterns for Solution Architects: Learn methods and principles of resilient distributed application architectures from hybrid cloud to far edge

✍ Scribed by Ashok Iyengar and Joseph Pearson

• Publisher: Packt Publishing Pvt ltd
• Year: 2024
• Tongue: English
• Leaves: 290
• Category: Library

✦ Synopsis

Enriched with insights from a hyperscaler’s perspective, Edge Computing Patterns for Solution Architects will prepare you for seamless collaboration with communication service providers (CSPs) and device manufacturers and help you in making the pivotal choice between cloud-out and edge-in approaches.
This book presents industry-specific use cases that shape tailored edge solutions, addressing non-functional requirements to unlock the potential of standard edge components. As you progress, you’ll navigate the archetypes of edge solution architecture from the basics to network edge and end-to-end configurations. You’ll also discover the weight of data and the power of automation for scale and immerse yourself in the edge mantra of low latency and high bandwidth, absorbing invaluable do's and don'ts from real-world experiences. Recommended practices, honed through practical insights, have also been added to guide you in mastering the dynamic realm of edge computing.

By the end of this book, you'll have built a comprehensive understanding of edge concepts and terminology and be ready to traverse the evolving edge computing landscape.

✦ Table of Contents

Edge Computing Patterns for Solution Architects
Contributors
About the authors
About the reviewer
Preface
Who this book is for
What this book covers
Images used in the book
Conventions used
Get in touch
Share Your Thoughts
Download a free PDF copy of this book
Part 1:Overview of Edge Computing as a Problem Space
1
Our View of Edge Computing
Suggested pre-reading material
Speaking like an edge native
What is the edge?
Are the edge and the cloud extremes of the same thing?
How does edge computing bring value, and why now?
Which edge? Categorizing edges
The user edge – field-deployed compute
The SP edge – regional compute
Your computer or mine? Tactics for service deployment
Edge computing doesn’t require dedicated resources
Single device, running multiple applications simultaneously
Single device, alternating applications by schedule or purpose
Hosted edge infrastructure – applications on shared compute
Cloud-out versus edge-in
Looking deeper at cloud-out architectures
Delving into edge-in architectures
Introducing archetype patterns
What is an archetype?
The days of software creation
Deploying archetype patterns
Summary
2
Edge Architectural Components
Edge components
Functional requirements
Sensing
Inferencing
Analytics
Data
Non-functional requirements
Security
Service management and operations
Edge use cases and patterns
Edge device specifications and protocols
Architectural decisions
Grouping edge ADs
Cloud
Network
Server/cluster
Device
Summary
Part 2: Solution Architecture Archetypes in Context
3
Core Edge Architecture
Suggested pre-reading material
What is legacy IoT architecture?
A bit of history
Purpose and promise
Fundamental drawbacks
Device configuration
Rationale
Architectural element categories
Edge devices versus edge hub
Reviewing the pattern
Self-propelled inspection robot example
Containers
Disconnected operations
Summary
4
Network Edge Architecture
Definitions
NFV
NFV considerations
SDN
VNF, NFV, SDN, and edge computing
Underlay and overlay networks
Network traffic management
MEC
Network edge architecture
RAN
CSPs and hyperscalers
Sample architectures
Manufacturing scenario
Healthcare scenario
Campus network scenario
Summary
5
End-to-End Edge Architecture
IT and OT convergence
AI and edge computing
Industrial edge scenario
Manufacturing scenario
Retail edge scenario
Retail store scenario
Network slicing
Example scenario
Edge reference architecture
The cloud
The network
The edge
Edge and distributed cloud computing
Distributed cloud computing
The scenario
Summary
Part 3: Related Considerations and Concluding Thoughts
6
Data Has Weight and Inertia
Suggested pre-reading material
Data encryption
Motivations for encrypting data
Protecting data without making it difficult to use
Ensuring that data modifications are noticeable
Data storage and management
Strategies for defining and enforcing data policies
Usage options ranging from real to synthetic data
Rules of thumb for retaining data, or not
Using data to build machine learning (ML) models
The promise of foundation models
How small and efficient can we make models?
Customizing existing models for each deployment
Using general-purpose platforms rather than single-purpose applications
Connectivity and the data plane
Optimizing data availability without connectivity
Aggregating data versus keeping it distributed
Migrating data files automatically
Summary
7
Automate to Achieve Scale
Automating service delivery
DevOps
Infrastructure as code
Extending automation to the edge
Developing edge applications
Scalability with automation
Prepping an edge device
Prepping an edge cluster
Operational security
Limiting physical access
Limiting connectivity
Trusted hardware and provisioning
Trusted data
Trusted compute
Tactical Edge
Automation with AI
LLMs and generative AI
Using AI in automation
Summary
8
Monitoring and Observability
Monitoring and observability
How monitoring works
How observability works
How network observability works
Measuring to improve
Network observability example
What to measure
Real user monitoring
Network performance management
Anomaly detection
Capacity
Business outcomes
Improving edge solution
Monitoring challenges at the edge
Configuration changes at the edge
Edge application monitoring
Personas
Summary
9
Connect Judiciously but Thoughtlessly
Suggested pre-reading material
Declarative versus imperative configuration
Comparing the two approaches
What slows down application deployment on the edge?
Solutioning edge-connected networks and applications
Zero Trust or as close as you can get
Managing secrets on the edge
Zero Trust architectures in edge computing
Secure access service edge
Overlay, underlay, and shared responsibilities
The network underlay
The network overlay
Zero Trust Network Access
End-to-end encryption
Application-centric networking
Summary
10
Open Source Software Can Benefit You
Suggested pre-reading material
Open source and edge computing
Edge computing and OSS are intertwined
Do you really need to create that component?
Creating and supporting an open source program office (OSPO)
A software bill of materials is your friend
Using SBOMs to track software dependencies
The characteristics of a mature OSS project
How to nurture and assist projects you rely on
Responses to projects that stray from their mission
Common objections
Recommendations for contributing code
Let the cat out of the bag (Successfully open source your code and documentation)
Five options for open sourcing
What to open source
Summary
11
Recommendations and Best Practices
Suggested pre-reading material
Edge-native best practices as an outgrowth of cloud native
Pulling can be more secure than pushing
Application dependency resolution approaches
Deployment models for distributed edge applications compared
Making antifragile applications
Defining the terms
What are your current areas of weakness or vulnerability?
Properties of antifragile architectures
An ounce of prevention...
When things go wrong
What to avoid
Anti-patterns
How to recover, gracefully or not
Summary
Index
Why subscribe?
Other Books You May Enjoy
Packt is searching for authors like you
Share Your Thoughts
Download a free PDF copy of this book