Background
Global factory owner that needs to remotely deploy and run machine models on-premise without internet connectivity due to advanced security architectures. Specific requirements in terms of cloud services and providers. Container-based deployments needed.
Prerequisites
A high number of factory sites. The factory owner has structuralized both local and overall architecture to meet future requirements for connected IT and OT systems. No connection between machine and internet is allowed. IoT system has to run behind a local firewall. Edge deployments can run on local PC:s or virtual machines on the factory site.
Scope
IoT behind factory firewalls
- All data communication has to be initiated from inside the firewall. No incoming request allowed.
Docker deployments requirements
- All services have to be distributed containerized.
Advanced Machine Models
- Run Machine specific Python models to analyze machine runtime behavior
- Real-time analysis to trigger actions in other machines (Machine Learning)
- Integrate into the global system remotely for further RnD and further machine learning objectives
Solution
Crosser Node deployed in the machine data layer (virtual machine):
- MQTT data communication via service bus to the internet access layer.
- Deployment of Python Machine Models for detecting anomalies within different use-cases
- Sending data outside the factory firewall to customer head quarters
Crosser Cloud:
- MQTT data communication to Crosser Cloud for updating node functionality
- Customer organization setup with roles
- Resources library setup
learn more about Crosser MLOps here →