Customers build a variety of streaming applications by using Azure Stream Analytics, such as IoT, connected car and automotive telematics, and smart elevators. This work requires processing of telemetry streams from each asset or source by itself—all in the same job, without merging the timeline from events that belong to different devices or sources. The reason is that their clocks might be drastically out of sync.
Substream support in Stream Analytics helps customers process telemetry streams with simple language constructs. A new keyword, OVER, extends the TIMESTAMP BY clause for this purpose.
For more details, see the TIMESTAMP BY documentation.