No no storage limits, just reach for the sky, with data virtualization in SQL MI
summary
Azure SQL Managed Instance queries data directly on Azure Blob/ADLS Gen2 (CSV, Parquet, JSON) via external tables or OPENROWSET, using managed identity auth, so cold data lives off-instance while queries join local and external sources. Practical savings: MI has a 16 TB local limit, but a demo shows scanning 4.7M remote rows reduced to ~23K with predicate pushdown, cutting query time from ~5s to near instant by using parquet and the Latin1_General_100_BIN2_UTF8 collation; partition elimination and file functions also work.
AI-written summary · about
chapters (11)
- 0:00Introduction and overview
- 2:04What is Azure SQL MI
- 5:13Why data virtualization
- 7:21How data virtualization works
- 11:17Demo: Data exploration
- 19:10Demo: Partitioned data access
- 23:12Performance optimization tips
- 29:13File functions and partition elimination
- 36:35Mixing local and remote data
- 41:28Roadmap and future plans
- 47:18Q&A and closing
related talks
52mAzure Sql DB Managed instance, a different sql server?
44mSQL Managed Instance: Fully managed SQL Server in Azure
51mFly at Mach-speed with Azure SQL Managed Instance- 51mFly at Mach-speed with Azure SQL Managed Instance
39mModernize your applications on Azure SQL Managed Instance Q&A
19mThere and back again: SQL Server 2022 - Managed Instance
44mThe Ultimate Freedom of Movement with SQL Server 2022 and Azure SQL Managed Instance
01h 01mLatest Innovations for SQL and Operational Databases for Developers