SQL Server 2025: The Developer’s Revolution

Description

Explore SQL Server 2025’s new dev features -JSON type, Regular Expressions, and Change Event Streaming for real-time pipelines. See how REST API and Data API Builder turn your data into secure, high-performance APIs. Packed with demos, this session is perfect for developers, DBAs, and architects

Key Takeaways

• Influence our SQL roadmap and ensure
• IT Director – DBA and BI Services at MSC
• Working with the SQL Server for 30 years
• Microsoft MVP - Data Platform
• Microsoft Certified Trainer
• Microsoft Certified: Fabric Analytics Engineer
• Former Azure Data Community Advisory Council

My Notes

Action Items

• [ ]

Resources & Links

Slides

SQL Server 2025:
The Developer’s
Revolution
Javier Villegas / Gonzalo Bissio
Sound off.
The mic is all yours.
Influence the product roadmap.
Join the Fabric User Panel
Join the SQL User Panel
Share your feedback directly with our
Fabric product group and researchers.
Influence our SQL roadmap and ensure
it meets your real-life needs
https://aka.ms/JoinFabricUserPanel
https://aka.ms/JoinSQLUserPanel
Javier Villegas
• IT Director – DBA and BI Services at MSC
• Working with the SQL Server for 30 years
• Microsoft MVP - Data Platform
• Microsoft Certified Trainer
• Microsoft Certified: Fabric Analytics Engineer
Associate
• Former Azure Data Community Advisory Council
• Technical Speaker
• Build, Ignite, PASS Data Community Summit, SQL
Saturdays, PASS Marathon, PASS Virtual Groups,
Microsoft AI+ Tour, Data Saturdays, Azure Global
Bootcamp, Group By , Power BI Summit,
DataPlatformGeeks & Fabric Conference
Gonzalo Bissio
• Head of cloud at Readymind
• Microsoft MVP - Data Platform
• Microsoft Certified Trainer
• Microsoft Certified: Fabric Analytics Engineer
Associate
• Technical Speaker
• Build, Ignite, PASS Data Community Summit, SQL
Saturdays, PASS Marathon, PASS Virtual Groups,
Microsoft AI+ Tour, Data Saturdays, Azure Global
Bootcamp, Group By , Power BI Summit,
DataPlatformGeeks & Fabric Conference
Generally Available
SQL Server 2025
The AI-ready enterprise database from ground to cloud
AI built -in
Built for
developers
Mission Critical
aka.ms/sqlserver2025
The AI-ready enterprise database from ground to cloud
Develop modern
data applications
AI built-in
Integrate your data
with Fabric
Secure by default
Connected
with Arc
Mission critical engine
Managed
DR
Assisted by
Copilots
Benchmark1
leader
Built for all
platforms
What problems are we trying to solve with AI?
Smarter searching on your existing text data
Bring in other documents/text for centralized vector searching
Provide building blocks - Intelligent assistants, RAG, AI Agents
Take advantage of AI in a secure and scalable fashion
Overcome complexity by using the familiar T-SQL language
SQL Server 2025 Vector Architecture
Locally hosted
REST HTTPS
T-SQL
Model
definition
Generate
Embeddings
T-SQL Prompt
AI Services and
Language
Extensions
Generate
Embedding
for prompt
Existing data
Cloud hosted
REST HTTPS
Create
vector
index
API_TYPE
Azure OpenAI
OpenAI
Ollama
ONNX Runtime*
Vector
and
Hybrid
Search
Network hosted
REST HTTPS
Built for developers
Develop modern data applications
JSON
Data API Builder
REST or GraphQL
There’s more…
GitHub Copilot for MSSQL
Change Event
Streaming*
JSON type, index,
and updated T-SQL
functions
Consume log change
events
T-SQL
REST API
RegEx and other
T-SQL functions
Connect your data
to any REST interface
with HTTPS
Microsoft Python Driver
for SQL Server

• Enabled by PREVIEW_FEATURES
  JSON
  < SQL Server 2025
  SQL Server 2025
  varchar/nvarchar column
  json data type – binary storage up to 2GB
  Any index that supports varchar/nvarchar
  .modify operator
  Various functions for JSON
  json index

• JSON_OBJECTAGG and JSON_ARRAYAGG
• JSON_CONTAINS
• JSON_QUERY WITH ARRAY WRAPPER
  Save space with compressed storage
  Less I/O and page reads
  REST in the engine?
  Off by default
  Securely avoid roundtrips to send data to a REST endpoint
  What is the REST protocol for your endpoint? curl.exe is your friend…so is Copilot
  EXECUTE @returnValue = sp_invoke_external_rest_endpoint
  [ @url = ] N'url’
  HTTPS local or remote
  [ , [ @payload = ] N'request_payload’ ]
  This is the crucial piece
  [ , [ @headers = ] N'http_headers_as_json_array' ]
  [ , [ @method = ] 'GET' | 'POST' | 'PUT' | 'PATCH' | 'DELETE' | 'HEAD' ]
  [ , [ @timeout = ] seconds ]
  [ , [ @credential = ] credential ]
  [ , @response OUTPUT ]
  API key, SAS token, managed identity
  NVARCHAR but often JSON
  HTTP_EXTERNAL_CONNECTION wait
  [ , [ @retry_count = ] # of retries if there are errors ]
  RegEx
  functions
  Some T-SQL Love
  Fuzzy string match
  functions
  UNISTR()
  PRODUCT()
  BASE64
  functions
  CURRENT_DATE
  returns DATE
  SUBSTRING()
  optional length
  ||
  string concat
  operator
  DATEADD()
  supports bigint
  The T-SQL RegEx library
  LIKE
  REGEXP_LIKE
  REPLACE
  REGEXP_REPLACE
  SUBSTR
  REGEXP_SUBSTR
  CHARINDEX, PATINDEX
  REGEXP_INSTR
  LEN and REPLACE
  REGEXP_COUNT
  CHARINDEX, PATINDEX,
  SUBSTR
  REGEXP_MATCHES
  CHARINDEX, REPLACE,
  SUBSTR
  REGEXP_SPLIT_TO_TABLE
  Change Event Streaming (CES)
  CES enables you to stream your data changes in near real-time
  from SQL data directly into Azure Event Hubs.
  Real-Time stream
  processing and Analytics
  Azure Event Hub
  Event Based Applications
  AI Agents
  Transaction Auditing
  Capturing changes with SQL Server
  Change Tracking (CT)
  Change Data Capture (CDC)
  Change Event Streaming (CES)
  Capture Source
  As part of committed transactions (sync)
  Committed transactions from transaction log
  (async). Requires SQL Server Agent
  Committed transactions from transaction
  log (async)
  Capture Destination
  System table (minimum)
  System table (extra logging required)
  Event stream (“IOless”)
  Data Captured
  Row identifier (DDL allowed)
  Row identifier and data
  Row identifier and data (DDL allowed)
  Capture Method
  Query system tables and user tables.
  Multiple destinations
  Query system tables. Multiple destinations
  Process event stream (“IOless”). Single
  destination
  Availability
  Immediate (pull-based)
  Near real-time (pull-based)
  Near real-time (push-based)
  SQL effects
  Transaction commit
  Transaction log truncation
  .
  Transaction log truncation
  Typical Use Cases
  Cache invalidation and sync
  Data warehousing and auditing
  Event-driven architectures – microservices
  integration, real-time analytics, cache sync,
  AI Agents
  MSSQL extension for VS Code
  Generally Available
  GitHub Copilot Ask and Agent mode
  Chat, inline suggestions, and slash commands
  Product Updates panel (new features, changes, resources)
  Public Preview
  View and Edit data
  Data-tier Application (DACPAC/BACPAC) export/import
  SQL Database Projects – Publish Dialog
  Drivers
  Generally Available
  Python Driver: mssql-python
  Latest versions of Microsoft.Data.SqlClient (MDS) and JDBC all
  support the new vector and JSON data types
  Support for vector in ODBC coming soon
  The SQL Server 2025 Platform Architecture
  Azure AI Agent
  Azure
  Microsoft Fabric
  Foundry Models
  Microsoft Entra
  Managed
  Identity
  Mirroring
  Azure Functions
  Azure EventHub
  Embeddings
  Change Event Streaming
  Azure Arc
  Data API Builder
  GraphQL
  Ai Agents
  Agent framework
  Managing the Server
  GitHub Copilot
  Local Dev Container
  SQL Server 2025
  Vectors
  JSON
  RegEx
  Concurrency
  Performance
  Security
  Availability Group
  Secondary Replica
  sp_invoke_external_rest_endpoint
  Embeddings
  Local
  REST service endpoint
  SSMS with
  Copilot
  VS Code mssql
  Reliable Failover
  Query Store
  Tuning
  Diagnostics
  Backups
  Agentic RAG
  AI Agents
  Local AI Models
  SSMS 22
  aka.ms/ssms
  Generally Available
  SQL Server 2025 support
  Fabric integration
  GitHub Copilot (Preview)
  Windows ARM64 support
  Migration Assistant
  Query Hint Recommendation tool
  New themes including dark
  JSON viewer
  Zoom results grid
  No pricing or licensing changes
  SQL Server 2025 editions
  The AI-ready enterprise database from ground to cloud
  Express
  Standard
  Enterprise
  Free, entry-level database for small AI, web
  and mobile apps
  Full featured database with extensive scale for
  mid-tier applications
  Highest performance and scaling for businesscritical workloads
  Feature highlights
  Feature highlights
  Feature highlights
• Express features
• Express and Standard features
  • Up to 1410 MBs of memory
  • Up to 32 cores of CPU NEW
  • Highest core count supported
  • Built-in vector search and semantic search NEW
  • Up to 256 GB of memory NEW
  • Highest memory supported
  • Native JSON type support NEW
  • Optimized locking NEW
  • Built-in Intelligent Query Processing
  • External REST endpoint invocation NEW
  • Resource Governor
  • Online and Resumable Index Operations
  • Change event streaming NEW
  • ZSTD Backup Compression Algorithm
  • Microsoft Entra ID authentication NEW
  • Basic Availability Groups
  • Query Store for readable secondaries
  • Mirroring in Fabric NEW
  • Azure SQL Managed Instance link
  • Persisted Statistics for Secondary Replicas (Query store)
  • 50 GB maximum relational databases size NEW
  • Up to 4 cores of CPU
  NEW
  • Availability Groups: including distributed and contained
  NEW
  • Accelerated Database Recovery
  • Backups to S3-compatible object storage
  • 524 PB maximum relational databases size
  Developer
  Free to use for dev/test in non-production environments
  Standard Developer NEW
  Enterprise Developer