Guide to Deploy MATLAB on Akash

The Akash Network provides a decentralized cloud platform where you can deploy containerized applications. This guide will walk you through deploying MATLAB on Akash using the official MathWorks MATLAB Docker image.

Prerequisites

1. MATLAB License:

   • MATLAB requires a valid license. You will need your license server information or network license key for deployment.

2. Akash CLI Setup:

   • Install and configure the Akash CLI.
   • Fund your Akash wallet with AKT tokens for deployment.

3. Docker Knowledge:

   • Familiarity with Docker containers is recommended.

4. Access to MATLAB Docker Image:

   • MathWorks provides the MATLAB Docker image on Docker Hub.
   • You need Docker Hub credentials to access the image.

Step 1: Pull the MATLAB Docker Image

1. Log in to Docker Hub with your credentials:

   docker login

2. Pull the MATLAB Docker image:

   docker pull mathworks/matlab:latest

Step 2: Prepare the SDL File for Akash Deployment

1. Create an SDL file (deploy.yml) for MATLAB. Use the following template:

   version: "2.0"
   
   services:
     matlab:
       image: mathworks/matlab:latest
       args:
         - --licenseserver <LICENSE_SERVER_HOST>:<LICENSE_SERVER_PORT>
       env:
         - MATHWORKS_LICENSE_FILE=<LICENSE_SERVER_HOST>:<LICENSE_SERVER_PORT>
       expose:
         - port: 8888
           as: 8888
           to:
             - global
       resources:
         cpu:
           units: 500
         memory:
           size: 512Mi
         storage:
           size: 1Gi
   
   profiles:
     compute:
       matlab:
         resources:
           cpu:
             units: 500
           memory:
             size: 512Mi
           storage:
             size: 1Gi
   
   deployment:
     matlab:
       matlab:
         profile: matlab
         count: 1
   ```
   
   - Replace `<LICENSE_SERVER_HOST>` and `<LICENSE_SERVER_PORT>` with your MATLAB license server details.

Step 3: Deploy to Akash

1. Create a Deployment:

   • Submit your SDL file to Akash:

     provider-services tx deployment create deploy.yml --from <your_account>

2. Find a Provider:

   • Query for providers willing to host your deployment:

     provider-services query market bid list --owner <your_account>

3. Accept a Lease:

   • Accept a lease from a provider:

     provider-services tx market lease create --from <your_account> --dseq <deployment_sequence>

4. Access Your Deployment:

   • Get the public endpoint of your deployment:

     provider-services query market lease get --dseq <deployment_sequence>

   • Use the endpoint to access your MATLAB application.

Step 4: Verify MATLAB Deployment

1. Open a browser and navigate to the public endpoint (e.g., http://<endpoint>:8888).
2. Verify that MATLAB is running and functional.

Step 5: Monitor and Manage Deployment

• Use the Akash CLI to check logs and manage your deployment:

  provider-services logs --dseq <deployment_sequence> --from <your_account>

• If you need to update or terminate the deployment:

  provider-services tx deployment close --dseq <deployment_sequence> --from <your_account>

Notes

• Firewall Configuration: Ensure the port 8888 is accessible globally or restrict it based on your needs.

• MATLAB Licensing: MATLAB requires a valid license at runtime. The deployment will fail if the license is invalid or inaccessible.

• Resource Optimization: Adjust CPU, memory, and storage in the resources section of the SDL file based on your workload.

This guide provides a basic setup to deploy MATLAB on Akash. Depending on your specific use case, you can further customize the deployment by modifying the SDL file or integrating additional services.