Edge Computing Systems

Building Intelligent Applications Across Devices, Edge, and Cloud

Zheng Song · English manuscript v0.4.2

This web edition contains the complete English manuscript: all fourteen chapters, five appendices, references, and image credits. Chapter and figure numbers follow the manuscript.

About this book

• Preface
• About this web edition

Part I. Applications, Devices, Communication, and Digital Twins

• Chapter 1. Introduction to Edge Computing
  • 1.1 From Connected Devices to Edge Computing
  • 1.2 Why Edge Computing?
  • 1.3 Devices, Edge Servers, and Cloud Resources
  • 1.4 Where Is the Edge? Common Deployments
  • 1.5 The Edge–Cloud Continuum
  • 1.6 From Application Needs to Architecture
  • 1.7 History and Rapid Development of Edge Computing
  • 1.8 Challenges and the Road Ahead
  • Chapter summary
  • Exercises
• Chapter 2. An Agentic IoT Application at the Edge
  • 2.1 From a Fixed Dashboard to On-Demand Analysis
  • 2.2 Why a Language Model, and Why at the Edge?
  • 2.3 The Components and Their Responsibilities
  • 2.4 Data and Tool Contracts That Support Planning
  • 2.5 Follow a Request from Goal to Result
  • 2.6 Requirements, Initial Deployment, and the Work Ahead
  • Chapter Summary
  • Exercises
• Chapter 3. IoT Devices and Sensing Systems
  • 3.1 A map of the physical edge
  • 3.2 Representative device classes
  • 3.3 From physical quantity to application observation
  • 3.4 Common sensing modalities
  • 3.5 Sampling, resolution, calibration, and uncertainty
  • 3.6 Connecting sensors to computing devices
  • 3.7 Smartphone sensing as an edge-systems problem
  • 3.8 Wearables and participatory sensing
  • 3.9 Vehicle and robot sensing
  • 3.10 Fusion, coordinate frames, and time alignment
  • 3.11 Data and energy budgets across modalities
  • 3.12 Laboratory and exercises
  • Chapter perspective
• Chapter 4. Communication and Networking for IoT and Edge Systems
  • 4.1 A network is a set of responsibilities
  • 4.2 Keep the layers separate
  • 4.3 Local access: wired networks, Wi-Fi, and Bluetooth LE
  • 4.4 Low-power meshes: IEEE 802.15.4, Zigbee, and Thread
  • 4.5 Wide-area IoT: LoRaWAN and cellular access
  • 4.6 Network, transport, discovery, and protection
  • 4.7 Application interaction: HTTP, MQTT, and CoAP
  • 4.8 Robotics, industry, and semantic interoperability
  • 4.9 Work a communication budget before choosing a stack
  • 4.10 Compare complete designs across applications
  • 4.11 From IoT Data to Digital Twins
  • 4.12 Laboratory: inspect an exchange and its boundaries
  • 4.13 Exercises
  • Chapter perspective

Part II. Building and Operating Edge Systems

• Chapter 5. Running and Deploying Applications on Edge Nodes
  • 5.1 From a working prototype to an edge deployment
  • 5.2 Packaging and isolation: sharing a machine without sharing everything
  • 5.3 Cross-platform and remote deployment
  • 5.4 Organizing execution on limited resources
  • Chapter summary
  • Exercises
• Chapter 6. Scaling Edge Applications and Elastic Execution
  • 6.1 When one node is no longer enough
  • 6.2 Replicas need a stable entry point
  • 6.3 Managing replicas with Docker Swarm and Kubernetes
  • 6.4 From fixed replicas to elasticity
  • 6.5 Serverless: let the platform manage execution on demand
  • 6.6 Measure the expansion before claiming success
  • Chapter summary
  • Exercises
• Chapter 7. Computation Offloading and Device–Edge–Cloud Collaboration
  • 7.1 From similar workers to different execution locations
  • 7.2 Follow one computation from device to remote service
  • 7.3 Split responsibilities, not just execution time
  • 7.4 Where data lives, and what consistency means to the application
  • 7.5 Disconnection and recovery are part of collaboration
  • 7.6 Build and compare a small continuum application
  • Chapter summary
  • Exercises
• Chapter 8. Scheduling and Optimizing Edge Systems
  • 8.1 Define the decision before choosing a policy
  • 8.2 Locate the bottleneck with an application trace
  • 8.3 Arrange execution when jobs compete
  • 8.4 Choose placement, routing, and replica count together
  • 8.5 Adapt without spending more than the change saves
  • 8.6 Compare policies with a reproducible experiment
  • Chapter summary
  • Exercises
• Chapter 9. Commercial Edge Platforms: From Mechanisms to Deployment
  • 9.1 Read a platform through the system it must support
  • 9.2 Greengrass and IoT Core: manage software on devices you already have
  • 9.3 Wavelength: shared compute in a communications provider’s network
  • 9.4 Local Zones: place selected services near a region’s users
  • 9.5 Outposts: bring managed infrastructure to the site
  • 9.6 CloudFront functions and Lambda@Edge: execute within a delivery path
  • 9.7 Cloudflare Workers: a second managed runtime perspective
  • 9.8 Combine platforms according to requirements, not a universal price rule
  • 9.9 Turn architectural understanding into a bounded deployment
  • Chapter summary
  • Exercises

Part III. Edge AI

• Chapter 10. Running and Optimizing AI on Edge Devices
  • 10.1 Start with a prediction task, not a model name
  • 10.2 Understand just enough of the model to deploy it
  • 10.3 Build the first complete inference path
  • 10.4 Find the resource constraint before changing the model
  • 10.5 Choose a suitable model before compressing it
  • 10.6 Quantization: represent numbers with fewer bits
  • 10.7 Pruning: remove work the model may not need
  • 10.8 Knowledge distillation: teach a deployable student
  • 10.9 Optimize execution without changing the task
  • 10.10 Evaluate a deployment and keep it useful over time
  • 10.11 A reproducible small-model study
  • Chapter summary
  • Exercises
• Chapter 11. Collaborative Edge Intelligence: From Perception Models to Language Models
  • 11.1 Start with the reason for collaboration
  • 11.2 Spend extra computation only when it is useful
  • 11.3 Split one inference computation across nodes
  • 11.4 Collaborate on evidence, not only on compute
  • 11.5 Add language understanding without hiding the system
  • 11.6 Understand language-model latency and memory
  • 11.7 Choose a model for the request
  • 11.8 Collaborate while generating tokens
  • 11.9 Collaborate on subtasks and evidence
  • 11.10 Update future models without confusing training and inference
  • 11.11 Build and evaluate a bounded collaboration
  • Chapter summary
  • Exercises
• Chapter 12. Agentic AI and Tool-Driven Applications at the Edge
  • 12.1 Start with a goal that changes the work
  • 12.2 Separate the model, runtime, tools, and data
  • 12.3 Ground requests in local data and documents
  • 12.4 Turn a proposed plan into real tool execution
  • 12.5 Generate analysis programs without granting unrestricted execution
  • 12.6 Follow a changing request through the system
  • 12.7 Budget the whole task, not each individual call
  • 12.8 Place planning, data access, and analysis across the continuum
  • 12.9 Recover tasks without confusing a retry with new work
  • 12.10 Build the first tool-driven implementation
  • 12.11 Evaluate flexibility, correctness, and resource use together
  • Chapter summary
  • Exercises

Part IV. Security and Multi-Party Trust

• Chapter 13. Security, Privacy, and Safety in Edge Systems
  • 13.1 Start with the operation and its trust boundaries
  • 13.2 Establish identity, then grant limited authority
  • 13.3 Protect messages without confusing protection with truth
  • 13.4 Keep software trustworthy throughout its lifecycle
  • 13.5 Share edge resources without sharing authority
  • 13.6 Treat privacy as a property of the complete data path
  • 13.7 Authorize a request without confusing it with a safe action
  • 13.8 Use local intelligence for protection, and measure its mistakes
  • 13.9 Build an enforcement path and test what it cannot prove
  • 13.S Chapter summary
  • 13.E Exercises
• Chapter 14. Multi-Party Trust, Blockchains, and Distributed Ledgers
  • 14.1 Specify the disagreement before selecting a ledger
  • 14.2 Build a verifiable statement
  • 14.3 From individual records to a protected history
  • 14.4 Understand ledger roles and the path to acceptance
  • 14.5 Place the ledger on the right path of the continuum
  • 14.6 Use shared rules without pretending to automate the physical world
  • 14.7 A real IoT case: the Distributed Compliance Ledger
  • 14.8 Correct records, revoke authority, and handle offline readers
  • 14.9 Compare architectures and test their limits
  • 14.S Chapter summary
  • 14.E Exercises

Appendices and references

• Appendix A. Laboratory and Project Guide
  • A.1 Choose a laboratory by its learning objective
  • A.2 Observation and deployment foundations
  • A.3 Digital twins: retain observation meaning
  • A.4 Edge AI: infer first, then compare changes
  • A.5 Collaboration: preserve the comparison boundary
  • A.6 Agents: distinguish proposal, execution and evidence
  • A.7 Trust: test what each mechanism actually establishes
  • A.8 A course project beyond the examples
• Appendix B. Optional Extension: Learning and Adaptation Across Devices
  • B.1 Adaptation without training
  • B.2 Federated learning as a coordination pattern
  • B.3 Non-identical data and partial participation
  • B.4 Privacy and security do not follow automatically
  • B.5 Suggested extension exercise
• Appendix C. Selected Exercise Solutions
  • C.1 Chapter 1
  • C.2 Chapter 2
  • C.3 Chapter 3
  • C.4 Chapter 4
  • C.5 Chapter 5
  • C.6 Chapter 6
  • C.7 Chapter 7
  • C.8 Chapter 8
  • C.9 Chapter 9
  • C.10 Chapter 10
  • C.11 Chapter 11
  • C.12 Chapter 12
  • C.13 Chapter 13
  • C.14 Chapter 14
• Appendix D. Glossary and Notation
  • D.1 Core vocabulary
  • D.2 Units and recurring notation
  • D.3 Additional device, networking, and stream-processing terms
  • D.4 Agentic application vocabulary
  • D.5 AI, digital-twin and trust terminology
• Appendix E. Evidence and Reproducibility
  • E.1 Distinguish the kind of evidence
  • E.2 Reproduce without rewriting the record
  • E.3 Version models, policies and meaning
  • E.4 Validate the target environment
  • E.5 Read photographs and diagrams accurately
• References
  • Photograph and product-image credits