Another scenario is to use Azure for workloads that are only temporarily needed; for example, science calculations that need a high amount of calculation capacity and where you have a high investment in the hardware.
So, what do you do? Normally, you create a VM ARM template for those machines. Then, you create a task that triggers your VM deployment and the transfer of the raw data into Azure. The following diagram shows a draft of the workflow:
The deployed Azure VMs run the necessary calculations and deliver the result to an Azure Storage Account. The following diagram shows the process:
As soon as the calculations are finished, the automation task will shut down and delete the Azure VMs. Now, the final data can be downloaded, or you can use other Azure services such as Power BI Embedded or App Services to present or work further on it. The following diagram shows the workflow:
