Microservices and API Reference Architecture for Manufacturing Industry


 Manufacturing Production Functions

Realize Products

o  Design Product, Engineer Manufacture of Product, Provide Production Resources, Produce Products

Engineer Manufacture of Product

o  Define Engineering Problem, Define Production Processes, Design Production System, Develop Implementation Plan

Define Production Processes

o  Specify Process Requirements, Specify Materials Requirements, Specify Process and Material Flows, Specify Process Details, Estimate Production Cost

Design Production System

o  Specify Production Equipment, Specify Instrumentation & Control Systems, Specify Support Systems, Develop Facility Layout, Integrate & Test System Designs

Specify Instrumentation & Control Systems

o  Identify Control Requirements, Identify Instrumentation Requirements, Identify Communications Requirements, Integrate System Specifications

Develop Implementation Plan

o  Identify and Evaluate Supply Sources, Decide Make/Buy for Parts, Develop Facilities Plan, Develop Manufacturing Plan,

Provide Production Resources

o  Develop Capacity Plans, Create Supply Sources, Acquire Major Resources, Manage Plant Resources

Produce Products

o  Plan Production, Manage Materials, Schedule Jobs, Perform Jobs,

Perform Jobs

o  Direct Personnel and Machines, Control and Monitor Jobs, Coordinate Equipment Groups, Control Equipment



Events are communicated from the production machines, smart tools, and devices via:

o  Shop-floor protocols (OPC, OPC UA, MQTT)

o  TCP/IP or via field bus protocols (MODBUS or ProfiNet)


Data Formats

Device and machine data are typically represented as JSON or XML


Success Patterns

Production devices and machines are typically managed by DCS/SCADA systems

o  Can be integrated by industry protocols such as Profibus, OPC, and OPC-UA

Factory data collection, normalization, cleansing and stream processing support analytics (and production visibility) and actions being triggered by business rules

o  E.g., execute a workflow or feedback to the production cell (in the form of setpoint adjustments or commands) to dynamically reconfigure the manufacturing process

o  Likewise for the enterprise, with scope including all factories, systems, locations, etc

Manufacturing plants must be able to operate as a stand-alone unit from the enterprise

o  Some capabilities must reside in both the plant and the enterprise

o  Devices/machine data can be communicated up through layers (filtered/aggregated)


Architecture Diagram

No alt text provided for this image

Domain Map Diagram

No alt text provided for this image

Domain Map Outline

No alt text provided for this image

Domain IT Requirements

Comprehensive integration: Pervasive interoperability improves resource/equipment efficiency, predictive maintenance, and product/process quality

o  Vertical (machines to cloud): From the operational technology layer to the information technology layer

o  Horizontal (among supply networks)

o  Throughout the product lifecycle

Data locality, privacy, and security (e.g., production data is not allowed to leave the factory)

Secure data access

Minimal technology gaps and complexity: Facilitates lean manufacturing

Event-stream processing

Real-time: No timing discrepancies between the operational technology layer (millisecond or even nanoseconds) and the information technology layer (sub-second and higher)

Low-latency event notification and reaction

Proactive monitoring and validation interconnected systems and applications

Auditing and compliance management: Reduce and control risks

Disconnected operation: In case central IT or cloud infrastructure is not available


Main theme: Service-orientation

o  Digitalization and integration of manufacturing resources as on-demand services

§ Facilitates value network integration and collaboration as well as plug-and-produce shop-floor systems

High-priority actions

o  Model and composition of micro-services

o  Optimize the topology of interactions between micro-services, smart devices and humans


o  Minimize (slow, error-prone) manual tasks

§ E.g., via RPA (robotic process automation), machine learning (ML) and artificial intelligence (AI)



API-first approach: Maximizes flexibility, customizability, scalability, reliability, manageability and changeability

Control, clarity and continuous improvement at all levels

Real-time data capture, processing and publishing improves decision-making by everyone:

o  Operators on the shop floor at the point of execution

o  Supervisors managing their departments

o  Executives evaluating annual performance targets

Efficient and transparent business performance management

Rapid adaptability to changing business requirements

Reduced maintenance, simplified migrations and expedited troubleshooting


Appendix: References

RAMI4.0 (Reference Architectural Model Industrie 4.0)

o  https://ec.europa.eu/futurium/en/system/files/ged/a2-schweichhart-reference_architectural_model_industrie_4.0_rami_4.0.pdf

IIRA (Industrial Internet Reference Architecture)

o  https://www.iiconsortium.org/IIRA.htm

o  https://www.iiconsortium.org/IIC_PUB_G1_V1.80_2017-01-31.pdf

NIST Reference Architecture for Smart Manufacturing

o  https://www.nist.gov/publications/reference-architecture-smart-manufacturing-part-1-functional-models

IBM Industry 4.0

o  Based on IBM’s IoT Reference Architecture, IIRA and the Purdue model of ISA-95

o  Description: https://www.ibm.com/cloud/garage/architectures/iotArchitecture/industrie_40

§ Blog at https://www.ibm.com/blogs/bluemix/2017/04/iot-industrie-40-reference-architecture/



Jordan Braunstein, CTO