README.md

Figures

Every captioned figure in Azure SQL Hyperscale Unveiled is published here, organized by chapter and named by its book number: ch02-fig-06-… is Figure 2-6. Many figures are hard to read at print size; open the file here to zoom in. Figures are reproduced as printed.

Back to the repository README.

Book media

File	Description
Chapter flip-through GIF	A looping, silent tour of all eleven chapters. 640 x 640, about 69 seconds, under 5 MB.
4K chapter video	A silent book flip-through for presentations and event screens. 3840 x 2160, about 2 minutes 26 seconds.
Book cover	The high-resolution front cover, including the foreword credit. 2100 x 3000 PNG.

The first inner page of both animations credits the book's quote contributors, including Joey D'Antonio.

Microsoft Azure icons used in the media are Microsoft-owned and subject to the Azure architecture icon terms, not this repository's MIT license.

Chapter figures

Chapter	Figures
Chapter 1	2
Chapter 2	11
Chapter 3	12
Chapter 4	8
Chapter 5	5
Chapter 6	11
Chapter 7	4
Chapter 8	6
Chapter 9	5
Chapter 10	6
Chapter 11	3
Diagram sources (SVG)	Editable architecture diagrams

Chapter 1

Figure	Caption
1-1	The road to Azure SQL Hyperscale
1-2	Hyperscale is built for mission-critical applications

Chapter 2

Figure	Caption
2-1	Azure SQL in the Azure portal
2-2	Creating a Hyperscale database in the Azure portal
2-3	Files and scripts in VS Code to deploy Hyperscale
2-4	Using AI assistance to make decisions like Zone Redundancy
2-5	Confirming choices before deploy
2-6	Hyperscale deployment confirmation
2-7	Connecting with the MSSQL extension
2-8	The Azure Activity log for the Hyperscale deployment
2-9	Using SSMS with Azure SQL Hyperscale
2-10	Using GHCP with your schema
2-11	Using GHCP with SSMS to explore data

Chapter 3

Figure	Caption
3-1	The Socrates architecture
3-2	The modern Azure SQL Hyperscale architecture
3-3	Looking at Hyperscale tiers
3-4	The Primary Compute in Hyperscale
3-5	Limits for Hyperscale Primary Compute
3-6	Tempdb files on Primary Compute
3-7	The Log Service
3-8	Page Servers
3-9	Data file storage in Hyperscale
3-10	Hyperscale replicas
3-11	List of Hyperscale replicas
3-12	End-to-end flows in Hyperscale

Chapter 4

Figure	Caption
4-1	The digital chart application
4-2	Top of patient digital chart
4-3	Bottom of the patient chart
4-4	Application architecture
4-5	Application source code in the GitHub repo for the book
4-6	Building the application with GitHub Copilot in VS Code
4-7	How to chat with your schema with GitHub Copilot
4-8	From the note using RegEx

Chapter 5

Figure	Caption
5-1	Patient Research in the application using vector search
5-2	Clinical Assistance in the digital chart
5-3	Application architecture with AI
5-4	The Hyperscale vector architecture
5-5	Hyperscale Agentic RAG

Chapter 6

Figure	Caption
6-1	LLMs and tools
6-2	The Model Context Protocol (MCP) — PNG · SVG
6-3	Data API Builder (DAB)
6-4	Using the SQL MCP Server with DAB
6-5	The Ward General digital chart application with AI agents
6-6	The Agent chat feature
6-7	The Agent side chat interface
6-8	Shift handoff candidates
6-9	Shift handoff approval
6-10	Shift handoff recommendations
6-11	The Agent flow

Chapter 7

Figure	Caption
7-1	Azure Monitor metrics in the portal for Hyperscale
7-2	CPU spike for Hyperscale
7-3	An Azure Copilot prompt for availability
7-4	Watching live data with XEvent

Chapter 8

Figure	Caption
8-1	The WardGeneral digital chart application with extra security
8-2	Choosing a provider in the digital chart application
8-3	Shift handoff with RLS applied
8-4	A private link for Hyperscale
8-5	The flow of Always Encrypted
8-6	Microsoft Defender for Cloud

Chapter 9

Figure	Caption
9-1	The Hyperscale HA architecture
9-2	The Hyperscale Zone Redundancy architecture
9-3	Hyperscale Geo-Replication architecture
9-4	The Hyperscale Failover Group architecture
9-5	A Hyperscale Geo-Replica

Chapter 10

Figure	Caption
10-1	Scaling the compute of the wardgeneral database
10-2	Hands-free data growth and IO for Hyperscale
10-3	The Hyperscale Elastic Pool architecture
10-4	JSON index seek in Hyperscale
10-5	Tempdb spill due to underestimated memory grant
10-6	Automatic Index Compaction in Hyperscale

Chapter 11

Figure	Caption
11-1	Why Azure is the best cloud for Hyperscale
11-2	The Azure AI Plane, Microsoft Foundry
11-3	Using skills to migrate to Hyperscale with DMS

Diagram sources (SVG)

The architecture diagrams are also published as SVG — the source several of the printed figures were drawn from. They stay crisp at any zoom, and you can open one and adapt it. Source file names use working numbers; the Book figure column gives the number printed in the book.

Chapter	Figure	Book figure	What it shows
1	Timeline	1-1	How Hyperscale got here, from SQL Azure to GA.
3	Architecture	3-2	The master figure: all four tiers and the paths between them.
3	Primary compute	3-4	The primary replica and its local RBPEX cache.
3	Log service	3-7	XLOG — the landing zone, the log broker, and who consumes the log.
3	Page servers	3-8	The 128 GB slices and the GetPage@LSN path.
3	File and storage	3-9	The durable tier, and why you pay on actual allocation.
3	Replicas	3-10	HA replicas, named replicas, and what attaches to what.
3	Transaction flow	3-12	One commit, followed across every tier.
4	Application architecture	4-4	The Ward General app, and the one connection it makes.
5	Application architecture	5-3	The same app once the engine itself calls out to models.
6	Application architecture	6-5	The same app once an agent, DAB and the SQL MCP server join it.
6	Agent prompt flows	6-11	How a shift handoff moves through the agent, the model, DAB, and SQL.
8	Security architecture	8-1	The same app with provider context, Row-Level Security, and the production Private Link path.

Chapters 4, 5, 6 and 8 are drawn on one shared grid on purpose. Put them side by side and the layout does not move — only what has been added to it.