MATLAB Production Server Workflow

The following figure illustrates the basic workflow to deploy MATLAB^® code using MATLAB Production Server™.

Workflow to deploy code to MATLAB Production Server

Create Deployable Archives

MATLAB application developers write MATLAB functions, and compile them into deployable archives using MATLAB Compiler SDK™. For more information, see Create Deployable Archive for MATLAB Production Server.

Deploy Archives to MATLAB Production Server

You can run MATLAB Production Server instances on-premises, in the cloud, or in Kubernetes^®.

Before you can deploy an archive to an on-premises instance, you need to create and configure a server instance. Configuring a server instance includes tasks such as configuring licenses and installing MATLAB Runtime. You can use the command line or the dashboard to manage the server. For more information, see Set Up MATLAB Production Server Using the Command Line and Set Up and Log In to MATLAB Production Server Dashboard.
Cloud deployments are pre-configured to use MATLAB Runtime. Cloud deployments offer the option of using a MATLAB Production Server license or pay-as-you-go for the services you use, where you do not require a license. For more information about MATLAB Production Server cloud offerings, see Cloud Deployment.
To configure a server in Kubernetes, you need to pull MATLAB Runtime container images from the MathWorks^® container registry and configure licensing. For deployment details in Kubernetes, see the MATLAB Production Server in Kubernetes repository on GitHub^®.

After a server is set up, server administrators can deploy the archives into one or more instances of the MATLAB Production Server. For more information, see Deploy Archive to MATLAB Production Server.

Write Client applications to Invoke Deployed MATLAB Code

Application developers use MATLAB Production Server client APIs or the MATLAB Production Server RESTful API to invoke MATLAB code deployed on the server. You can write client code in .NET, Java^®, Python^®, C, MATLAB, and web-based languages such as JavaScript^®. For more information, see Client Programming.

If applicable, you can use application installers to distribute and install client applications on end-user computers.