Equipment Gateway
Auto-generate equipment APIs from a single YAML config
The Problem
Protocol Fragmentation
Manufacturing floors use a mix of Modbus TCP, OPC-UA, and MQTT devices from different vendors. Each protocol requires its own library, connection logic, and data parsing — making integration a recurring engineering bottleneck.
Manual API Development Per Equipment
Every new equipment model required developers to write custom REST endpoints, register mappings, and polling logic from scratch. Adding a single product type could take days of boilerplate coding and testing.
No Unified Monitoring
Equipment status was checked through vendor-specific tools or manual inspections. There was no single dashboard to visualize real-time health, detect anomalies, or track historical trends across all connected devices.
Custom backend code written for every new equipment model
Add a YAML file and REST APIs are auto-generated — zero backend code per product
Separate tools and scripts per communication protocol
Unified protocol adapter layer handling Modbus TCP, OPC-UA, and MQTT transparently
Equipment status checked manually or via vendor-specific software
Real-time monitoring dashboard with time-series visualization and alerting
Approach
Configuration-Driven Architecture
Instead of writing code for each equipment type, the gateway reads YAML configuration files that define equipment models — registers, data types, polling intervals, and API endpoints. A code-generation layer transforms these definitions into fully functional REST APIs at startup. Protocol adapters abstract away communication differences, so the same YAML schema works whether the device speaks Modbus TCP, OPC-UA, or MQTT. This approach turns equipment onboarding from a development task into a configuration task, enabling operations engineers to add new products without writing a single line of code.
BDD Pipeline Flow
Key Features
YAML-Driven Equipment Config
Define equipment models declaratively in YAML — register addresses, data types (INT16, FLOAT32, BOOL), scaling factors, and polling tiers. The gateway parses these definitions at startup and auto-generates typed REST endpoints, entity schemas, and polling schedules. No backend code changes needed to onboard a new product.
Multi-Protocol Support
Pluggable protocol adapter architecture supporting Modbus TCP, OPC-UA, and MQTT out of the box. Each adapter implements a common interface for read/write operations, so equipment definitions remain protocol-agnostic. Adding a new protocol means implementing one adapter — all existing YAML configs work automatically.
Auto-Generated REST API
REST endpoints are dynamically generated from YAML definitions — GET for current values, POST for write commands, and WebSocket streams for real-time updates. The Entity Store maintains in-memory state with change detection, serving cached responses in microseconds while 3-tier polling (100ms / 500ms / 1s) keeps data fresh from the equipment.
Real-time Equipment Monitoring Dashboard
Visual dashboard built with ReactFlow showing live equipment topology and real-time sensor values. Time-series data streams into InfluxDB for historical trend analysis, anomaly detection, and configurable alerting thresholds. Engineers can monitor entire production lines from a single screen.
Architecture
Results
New equipment types onboarded with YAML config only
3-tier polling with register auto-merging for efficiency
Modbus TCP, OPC-UA, and MQTT via pluggable adapters
Real-time equipment health tracking with alerting