United Manufacturing Hub with Apache Kafka

Overview

United Manufacturing Hub Classic leverages Apache Kafka as its high-throughput message broker, enabling scalable event streaming for manufacturing data across the Unified Namespace.

Integration Architecture

In the UMH Classic Kubernetes deployment, Kafka serves as the central message broker that:

• Buffers all incoming industrial data from edge connectors
• Enables real-time streaming between OT and IT systems
• Provides durability and replay capabilities for critical manufacturing data
• Scales horizontally to handle high-throughput production environments

Use Cases

• Streaming sensor data from hundreds of PLCs to analytics platforms
• Building event-driven MES integrations
• Enabling real-time OEE calculations across production lines
• Supporting predictive maintenance data pipelines

Tradeoffs & Considerations

• Operational complexity: Kafka requires ZooKeeper and careful cluster management
• Resource requirements: Higher memory/CPU needs compared to simpler MQTT brokers
• UMH Core evolution: Newer UMH Core uses Redpanda (Kafka-compatible but simpler)
• Worth it when: You need high throughput (100K+ msgs/sec) or have multiple consumer groups