NEW AZURE REFERENCE ARCHITECTURE: Real-time scoring of R machine learning models

This post has been republished via RSS; it originally appeared at: Azure Global articles.

First published on MSDN on Dec 13, 2018
We have a new AI Reference Architecture (on the Azure Architecture Center) from AzureCAT Data Scientist, Hong Ooi. It was edited by Nanette Ray and Mike Wasson. It was reviewed by George Iordanescu (also from AzureCAT AI).

• Real-time scoring of R machine learning models

This reference architecture shows how to implement a real-time (synchronous) prediction service in R using Microsoft Machine Learning Server running in Azure Kubernetes Service (AKS).



This architecture is intended to be generic and suited for any predictive model built in R that you want to deploy as a real-time service. Deploy this solution.

The Reference Architecture includes the following information:

1. Architecture - Explaining the different elements of the architectural diagram.
2. Performance Considerations - What to watch out for to maintain high levels of performance.
3. Security Considerations - Network encryption, authentication, authorization, and separating storage.
4. Monitoring and Logging - We recommend the Kubernetes dashboard and Azure Monitor Insights.
5. Cost Considerations - How pricing works with Machine Learning Server, licensing alternatives, and managing the compute resources.
6. Deploy - Our GitHub repo includes prerequisites, setup instructions, deployment and testing steps, as well as the various R files and template.

Head over to the Azure Architecture Center to learn more about this reference architecture, Real-time scoring of R machine learning models.

See Also

Additional related Reference Architectures:

• Batch scoring on Azure for deep learning models
• Real-time scoring of Python Scikit-Learn and deep learning models on Azure

Find all our Reference Architectures here.




AzureCAT Guidance 
"Hands-on solutions, with our heads in the Cloud!"