4. Infrastructure as a Service
Batch
Application
Logic Apps
Turnkey managed services
SQL
Database
Redis
Cache
NoSQL
Search
HDInsight
Stream
Analytics
Machine
Learning
Event
Hubs
Marketplace
Active
Directory
...
Cognitive
Services
Power BI
Embedded
API
Management
Media
Services
Key Vault
Platform as a Service
App
Service
Service
Fabric
Functions Logic Apps
Serverless
Mono-lithic application -> loosely coupled components
Events happening async outside client application
Build/use micro-functionality to react to a single event
Abstraction of servers, infrastructure and configuration of operating system
“Serverless Computing” doesn’t mean that there is no server.
But it means that as an user of the service, you don’t have to know that they are some, which characteristics they have, what is the configuration and so on.
You don’t have to handle all the technical and administrative aspects of managing servers (high availability, backup, monitoring, scaling, …) and you care only about the service.
Event-driven scale
Sacling of the service is done transparently for the developer, based on the real usage of it.
If your company is broadcasted on a TV show, it will scale up to serve the requests. If the day before, you need to call the function once, it will scale down without any configuration.
Sub-second billing
Obviously, as a cloud service, you only pay for what you use, but the granularity is so small that we can talk about sub-second billing. I’m sure you’ve already seen some servers used at 10%, 1% or even less of their capacity. With serverless, that’s over.
Already deployed function app. This is the Function App portal, inside the Azure portal.
Create a new function, custom, python HttpTrigger
Run a show the log
Copy the URL, paste it on PostMan, with a POST and paste the Json payload before modifying it and send
In the Integrate tab, explain the notions of Triggers, Inputs and Outputs
Triggers are the one which trigger the function but does not always bring data. (http brings data through the querystring and the body but a timer won’t bring anything useful)
Input can be used to configure connectors to Azure services or external services to grab data when a trigger has triggered the function
Outputs can be used to configure connectors that will be used for your code to output data easily, without writting all the code needed to connect to the service
Show the Advanced editor in the integrate tab and the files in the code editor to tell people that everything can be done in plain files and deployed through Git for example.
Picture of a flower farm.
When growing flowers at high scale like this one, you have to monitor a few things, and to be proactive to detect any problem a flower can have.
One way to do that is to check the PH, the acidity of the water you bring to the flowers and the acidity of the water coming out. And if there is a difference, that might mean that you have a problem.
A ph sensor sending data to Azure :
Azure IoT Hub for bi-derectional communication,
Then Azure Stream Analytics to do some calculation
Azure SQL database to Store
Power BI to display the data on a dashboard
And if the calculation shows a problem, an item is sent to a queue with Azure Event Hubs
This queue is used a a trigger to begin a Azure Functions, which has the code needed to send an alert to an operator on his phone