Main Content

Cloud Deployment

Deploy MATLAB® Production Server™ on Microsoft® Azure® and AWS®

You can deploy MATLAB Production Server in the cloud using Microsoft Azure and AWS. The elasticity of the cloud infrastructure combined with MATLAB Production Server enables your application to support many users simultaneously.

MATLAB Production Server on Azure uses Azure Resource Manager (ARM) templates to automate the deployment and configuration of virtual machines for the server and the provisioning of network and storage resources.

The Azure Marketplace provides two offerings for MATLAB Production Server—bring your own license (BYOL) and pay as you go (PAYG). To use the BYOL offering, you must have a MATLAB Production Server license. For deploying MATLAB Production Server (BYOL), see MATLAB Production Server (BYOL). The PAYG offering does not require you to have a MATLAB Production Server license. Your costs for the PAYG solution depend on the number of vCPUs that you provision in Azure. For deploying MATLAB Production Server (PAYG), see MATLAB Production Server (PAYG).

If you need to customize the ARM templates and automation scripts or use a version of MATLAB Production Server in the cloud that is older than R2020a, you can use the MATLAB Production Server Reference Architecture on Azure instead. The MATLAB Production Server Reference Architecture on Azure is available on GitHub®. For deploying the reference architecture, see MATLAB Production Server on Microsoft Azure.

The MATLAB Production Server Reference Architecture on AWS is available on GitHub. For deploying the reference architecture, see MATLAB Production Server on Amazon Web Services.

The following table shows a comparison of the MATLAB Production Server cloud offerings and when you might want to use them.

 

MATLAB Production Server

Cloud Reference Architecture

Azure Marketplace (BYOL)Azure Marketplace (PAYG)
Ability to customize, maximum flexibility  
Move existing on-premises workloads to cloud  
Long-term trials, proofs-of-concept  
Bursty workloads