Before configuring devices or writing rules, understanding these core ideas will save you time and prevent mistakes later. Everything in Proxus is built around two principles: a Single Source of Truth for all operational data, and a Decoupled Architecture that separates producers from consumers.
After reading this section, you will understand:
- How data is organized across the platform (Unified Namespace)
- Why gateways are stateless and what that means for operations
- How the Hub & Spoke topology scales from one site to hundreds
- What data quality metrics tell you about your connected devices
Foundational Pillars
Hub & Spoke Topology
One Central Server coordinates all Edge Gateways — including a real multi-site scenario.
Unified Namespace (UNS)
The ISA-95 hierarchy that organizes all industrial data into a shared, semantic model.
Stateless Edge Architecture
Why gateways hold no state and how that makes large deployments simpler.
Embedded MQTT Broker
Accept device telemetry locally without installing external broker software.
Data Quality & Metrics
Track freshness, accuracy, and consistency across every connected device.
Performance Benchmarks
Pipeline throughput results and how to use them for capacity planning.
The Proxus Philosophy
Single Source of Truth (SSoT)
All operational data flows into a Unified Namespace. Dashboards, AI models, rules, and ERP systems all read from the same hierarchy. Nobody needs to know device addresses — they subscribe to meaningful topic paths like v1/acme/istanbul/assembly/line01/robot_a/telemetry/temperature.
Why this matters: When you add a new consumer (a dashboard, an analytics tool, a cloud connector), it simply subscribes to the UNS. No new point-to-point integration is needed.
Decoupled Architecture
Producers (PLCs, sensors) and consumers (dashboards, cloud, ERP) never talk to each other directly. Proxus sits in the middle, which makes three things predictable:
- Scaling: Add more producers or consumers without changing the existing ones
- Protocol translation: A Siemens PLC and a Modbus sensor both publish to the same namespace
- Resilience: If a consumer goes down, producers are unaffected. If a producer stops, other producers continue normally.
Core Design Principles
| Principle | What It Means in Practice |
|---|---|
| Stateless Edge | Replace a broken gateway in minutes, not days. No manual config rebuild. |
| Standardization First | Data is normalized when it enters the platform, not when someone tries to use it. |
| Open Access | MQTT, REST, SQL — your data is accessible through standard protocols. No vendor lock-in. |
| Resilience by Design | Network failures are expected. Store & Forward protects data during outages. |
Suggested Reading Order
If you are new to these concepts, we recommend reading them in this order:
- Hub & Spoke Topology — Start with the overall architecture
- Unified Namespace — Then understand the data model
- Stateless Edge — Then learn how gateways work
- MQTT Broker — Then understand device connectivity
- Data Quality — Finally, learn how to monitor data health
Or, if you prefer a role-specific guided path, see Learning Paths.