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

TodayI’d like to tell you about some very cool intelligence features withinthe Azure SQL Database. Azure SQL Database technologies deliverintelligent capabilities through a range of built-in machine learningand adaptive technologies that monitor and manage performance andsecurity for you.

Using telemetry from millions of databases running in Azure over theyears, Microsoft has built this capability of training a trulyintelligent and autonomous database that gives you the ability to learnand adapt to your workload. This intelligent performance gives you thedeeper insight into database performance. Plus, it eliminates the hassleof making ongoing improvements, allowing you to focus more on drivingyour 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 – thisgives you the ability to expand on batch-mode memory grant feedback byadjusting memory grant sizes for both batch and row mode operators.
• Approximate query processing – this isdesigned to provide aggregations across large datasets whereresponsiveness is more critical than absolute precision, and it willreturn an approximate value with the focus on performance.
• Table variable deferred compilation – thisimproves plan quality and overall performance for queries, referencingtable variable by propagating cardinality estimates that are based onactual table variable row counts. In turn, this optimizes yourdownstream plan operations.

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

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

In short, Microsoft has built these intelligent features over yearsof machine learning and is applying it to all their Platform as aService, as well as some of their on-premises, offerings. These arereally cool features and we’ve got great response about them and howwell 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.