AI-102 Question 44
Single answerYou are deploying a custom machine learning model as a web service in Azure Machine Learning. The deployed service needs to be accessible via a REST API. What must you configure to ensure the service has a default endpoint for communication?
- A
Deploy the model to an Azure Kubernetes Service (AKS) cluster and expose it using an Ingress Controller.
- B
Enable Azure Front Door for the deployed service to set up a default endpoint.
- C
Register the deployed model as a web service in Azure Machine Learning, which automatically creates a default endpoint.
- D
Publish the service to the Azure Marketplace for external access.
Show answer and explanation
Correct answer: C
Explanation
When deploying a machine learning model in Azure Machine Learning, registering it as a web service automatically creates a REST API endpoint that acts as the default endpoint for communication. This is the simplest and most direct way to enable access to the service without additional configurations.
- A. Incorrect.
Deploying the model to AKS and exposing it using an Ingress Controller is related to advanced configurations for scaling and exposing services. However, it does not automatically configure a default endpoint for the service.
- B. Incorrect.
Azure Front Door is used to provide global load balancing and routing, but it is not required to set up the default endpoint for a deployed service in Azure Machine Learning.
- C. Correct.
Registering the deployed model as a web service in Azure Machine Learning automatically creates a default REST API endpoint that can be used to consume the service.
- D. Incorrect.
Publishing to the Azure Marketplace is for making the service publicly available for consumption by external users but does not inherently involve setting up a default endpoint.