11_fan-out-in.md

Lab 11 - Fan-out/Fan-in

Goal

The goal of this lab is to apply the fan-out/fan-in pattern in order to perform activity functions in parallel.

Steps

The first two activity functions in the NeoEventProcessingOrchestrator can be executed in parallel. We have to wait until the result from both functions are available before we construct the TorinoImpactRequest.

1. Change to return tasks

Open the NeoEventProcessingOrchestrator orchestrator function and locate the first two activity functions:

• Remove the await keyword before the context.CallActivity... methods.
• Rename the variables which captures the output from the activity calls and postfix them with the word task. In my case I have variables kineticEnergyTask and impactProbabilityTask.

2. Await both tasks

Before constructing the TorinoImpactRequest object, add this:

await Task.WhenAll(kineticEnergyTask, impactProbabilityTask);

This ensures that the orchestrator code will only continue executing when all the tasks have been completed.

You can access the kinectic energy and impact probability by using the Result property on the respective tasks.

3. Build & run locally

Now run/debug your local Function App again. No functional change should be visible but now the first two activities are executed in parallel.

Continue to the next lab to use suborchestrations.