Understanding Azure Stream Analytics

(NOTE: A more up to date definition is available at “Understanding Azure Stream Analytics” in the May 2022 Update.)

Azure Stream Analytics is a high-speed data stream event processing solution available as a cloud service and as on-premises software to collect data locally in Internet of Things (IoT) edge scenarios. It provides real-time data stream analysis and event handling, such as processing Web clickstreams and IoT sensors, using a SQL-based query language and JavaScript functions and aggregates.

The service will likely appeal to organizations with SQL Server skills looking for a managed platform for event processing that removes the need to learn alternative technologies, such as Hadoop-based languages like Apache ActiveMQ, Kafka, and Spark, or Amazon’s Kinesis. However, the service is less flexible than many alternatives, and it limits data inputs and outputs to select Azure services and data formats.

How It Works

The service is based on the following components:

Data ingestion. Stream Analytics can accept streaming data from several Azure sources and custom interfaces, including Azure Event Hubs, IoT Hub, Azure Storage blobs, and Azure Data Lake Store (ADLS) Gen 2. The service can also query additional reference data (such as equipment details) stored in Storage Blobs, ADLS Gen 2, and Azure SQL Database, which can be used to enhance queries. For example, a query can compare the results of analyzing equipment performance in the streaming data with equipment threshold limits from reference data to determine if an alert should be generated.