When I decided to build the OT Digital Twin (OTDT) for EAAAIW 2026, I didn't fully grasp what I was signing up for. Deploying IBM Maximo Application Suite (MAS) 9.1 isn't like spinning up a Vercel app. It requires a Red Hat OpenShift cluster, massive compute resources, and a deep understanding of enterprise architecture.
I had zero budget, no team, and exactly 3 weeks.
The Architecture
The goal was to run a robust MAS 9.1 deployment. This meant I needed:
- A stable OpenShift (ROKS) cluster.
- Sufficient worker nodes to handle the heavy MAS pods.
- Persistent storage for databases and message queues.
IBM Cloud
└── ROKS Cluster (Frankfurt)
├── Worker Node 1 → MAS Core
├── Worker Node 2 → Manage Component
└── Worker Node 3 → IoT / Monitor
The Struggle
The hardest part wasn't the Kubernetes YAMLs or the terminal commands. It was managing the resources. As a student, enterprise-grade cloud credits are hard to come by. I had to:
- Optimize every single pod's requests/limits.
- Tear down environments daily to save credits.
- Automate the deployment process so I could rebuild the cluster in under an hour.
Scripting the Deployment
I ended up writing a series of Bash scripts to automate the OpenShift CLI (oc) commands. This saved me hours of manual configuration.
#!/bin/bash
# A snippet from my deployment automation
echo "Applying MAS namespace..."
oc apply -f mas-namespace.yaml
echo "Setting up operator subscriptions..."
oc apply -f mas-operator.yaml
The Result
By the time EAAAIW 2026 arrived, OTDT was fully functional. A 6-node enterprise cluster humming along, demonstrating real-time OT monitoring. It taught me more about systems architecture than any class ever could.
The takeaway? Constraints breed creativity. If you don't have a team, build the automation to be your team.