A Look at Some of Azure SQL Database’s Intelligence Features

Today I’d like to tell you about some very cool intelligence features within the Azure SQL Database. Azure SQL Database technologies deliver intelligent capabilities through a range of built-in machine learning and adaptive technologies that monitor and manage performance and security for you.

Using telemetry from millions of databases running in Azure over the years, Microsoft has built this capability of training a truly intelligent and autonomous database that gives you the ability to learn and adapt to your workload. This intelligent performance gives you the deeper insight into database performance. Plus, it eliminates the hassle of making ongoing improvements, allowing you to focus more on driving your business and less on “chores”.

Features like query performance insights and automatic tuning continuously monitor database usage and detect disruptive events and then they take steps to improve performance.

Three examples of the intelligent performance that can collectively optimize your memory usage and improve overall query performance are things like:

  • Row mode memory grant feedback – this gives you the ability to expand on batch-mode memory grant feedback by adjusting memory grant sizes for both batch and row mode operators.
  • Approximate query processing – this is designed to provide aggregations across large datasets where responsiveness is more critical than absolute precision, and it will return an approximate value with the focus on performance.
  • Table variable deferred compilation – this improves plan quality and overall performance for queries, referencing table variable by propagating cardinality estimates that are based on actual table variable row counts. In turn, this optimizes your downstream plan operations.

Along all those features, Azure SQL Database intelligent protection allows you to efficiently and productively meet your data’s security and compliance requests by proactively monitoring for potential threats and vulnerabilities. You can flag things such as PII or a cross-scripting attack or something like that. There are detection mechanisms in there that can help you avoid these.

Through features like information protection, vulnerability assessment and threat detection, you can proactively discover and protect sensitive data, as well as uncover potential vulnerability and detect anomaly activities that could indicate a threat to your data.

In short, Microsoft has built these intelligent features over years of machine learning and is applying it to all their Platform as a Service, as well as some of their on-premises, offerings. These are really cool features and we’ve got great response about them and how well they work.

I recommend you give these features a try, but remember, always try them out in your test or dev environments prior to bringing them into production.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.