Skip to content

anibalvelarde/uns-patterns

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

UNS Patterns

Welcome to the UNS Patterns repository. This is a collection of ideas aimed at sparking discussions on designing Unified Namespace (UNS) topic structures for industrial and manufacturing operations. Drawing from ISA-95 Part II's core hierarchy, the focus here is on providing generic starting points that can be extended to fit specific industries or needs. I'm sharing these as a way to explore principles, not as definitive solutions—feedback and contributions are very welcome (actually encouraged) to refine them.

What is UNS?

A Unified Namespace (UNS) is an architectural pattern for organizing real-time data in industrial systems, often built on pub/sub protocols like MQTT. It creates a hierarchical, browsable structure (e.g., /Enterprise/Site/Area/Line/Machine/Metric) that mirrors organizational assets, enabling decoupled data flows between OT (Operational Technology) and IT systems. In back-end terms, think of it as an event-driven data fabric—prototype with C# (MQTTnet) or Python (paho-mqtt) clients publishing to a broker like HiveMQ on your K8s cluster, with RDBMS (e.g., Postgres) for historical persistence.

Origin Story...

A huge thank you to Walker Reynolds (President of 4.0 Solutions) for coining the term "Unified Namespace" and championing its adoption in industrial IIoT. His practical insights from real-world integrations have shaped much of the modern discussion around UNS as an event-driven architecture for decoupling OT/IT systems.

Brief History

Walker prototyped the concept as early as 2005, building his first implementation using Dynamic Data Exchange (DDE) in a salt mine to unify machine data without physical trips to control booths. The term itself was formalized around 2015, evolving with MQTT and Sparkplug to address data silos in Industry 4.0. For a deeper dive, check this masterclass interview where he covers origins and evolution: Unified Namespace for Industrial IoT: The Masterclass (YouTube). This repo builds on those foundations but tailored to specific use cases as a starting point.

Tie to ISA-95 Part II

This repo builds on ISA-95's equipment and activity models for a standardized foundation. Start with a core hierarchy like:

 Enterprise > Site > Area > Production Unit > Work Cell > Equipment, 

then layer in variations based on manufacturing types or methodologies. Extend via YAML templates in /patterns/ for your custom setups—e.g., map to gRPC services for backend integration or AWS IoT Core for cloud scaling.

Open to Questions or Interpretation

There are many other classification schemes that have not yet been considered. The idea of UNS does not necessarily force you to choose anything presented here. This is a repository for examples the community can puth forth and others can take a look, experiment, refine, propose changes or present a brand new idea. Other facets that could have been used here include (but not limited to):

Motivations

What classification ontologies should be used to classify UNS patterns? It is difficult to say since your approach may be very different depending on your industry, corporate philosophy, company culture, etc., etc.

To get this effort off the ground a decision was made to consider ISA 95 Part II, Manufacturing Process Types and Manufacturing Methodology Types. Are there other classification schemes that can be used? Absolutely! If you do not see one you'd like listed, feel free to submit a pull-request!

For now, we'll use the following:

Manufacturing Processes

These types classify operations by production style, influencing UNS depth and granularity. For instance, a Job Shop might emphasize flexible /JobID/ sub-levels, while Continuous ops focus on flow metrics. Use as a base to adapt your topic structures.

Type Definition & Key Characteristics Examples Advantages Disadvantages
Job Shop Low-volume, high-customization production using flexible workstations (not lines) for make-to-order (MTO) or small batches; each job is unique, with workflows rerouted as needed. Custom furniture, prototypes, aerospace parts, specialty tools High flexibility for bespoke items; scalable to discrete if demand grows Inefficient for high volumes; hard to automate fully due to variability
Batch Produces identical items in groups (batches) based on demand; involves setup/cleanup between batches, often for fluids/powders but adaptable; machines handle most work. Baking goods, pharmaceuticals, chemicals, paints Flexible batch sizing; good traceability and quality control; minimal downtime for switches Downtime for cleaning/setup; requires identical items per batch
Repetitive High-volume production of similar/identical items on dedicated assembly lines running 24/7; minimal changeovers, focused on efficiency for standardized products. Appliances, electronics, consumer goods like toys or electrical components High efficiency and scalability; low setup time for consistent runs Low flexibility; overproduction risk if demand drops
Discrete Assembly of distinct, countable items from components; lines allow frequent changeovers for product variations (e.g., sizes/styles); products can be disassembled/recycled. Automobiles, smartphones, furniture, airplanes Handles product diversity; efficient for mid-to-high volumes with variations Time-consuming setups for changes; more complex than repetitive
Continuous Non-stop flow of raw materials (gases, liquids, powders) through chemical/physical transformations; 24/7 operation with uniform output, no batches. Oil refining, metal smelting, paper production, food like peanut butter Extremely efficient for high volumes; minimal interruptions Inflexible for variations; high initial setup costs; faults propagate quickly
Additive (e.g., 3D Printing) Layer-by-layer building from digital designs; enables complex, customized low-volume parts; often seen as a supplement to traditional types. Medical devices, prototypes, aerospace components, custom visors Ultimate customization; reduces waste; innovative for complex geometries Slower for high volumes; material limitations; higher per-unit costs

Manufacturing Methodologies

These approaches guide process optimization and can overlay on types to refine UNS structures—e.g., add /Constraint/ levels for TOC or /Sprint/ for Agile.

Methodology Definition & Core Focus Key Tools/Principles Examples in Practice Pros Cons
Lean Manufacturing Eliminates waste (muda) in all forms (e.g., overproduction, waiting, defects) via continuous flow and value-stream focus; rooted in Toyota Production System (TPS). 5S (Sort/Set/Shine/Standardize/Sustain), Kaizen, JIT, Kanban, Value Stream Mapping Automotive (Toyota assembly lines), electronics (reducing inventory in PCB manufacturing) Reduces costs/inventory; boosts efficiency; adaptable to any scale Requires cultural buy-in; initial disruptions during implementation
Six Sigma Data-driven methodology to minimize process variability and defects (aiming for 3.4 DPMO); uses DMAIC (Define/Measure/Analyze/Improve/Control) cycle. Statistical tools (e.g., control charts, DOE), root cause analysis (fishbone diagrams), process mapping Pharma (defect-free drug batches), semiconductors (yield optimization) Quantifiable improvements; strong on quality metrics Data-heavy; can be rigid/overly analytical without Lean integration
Total Quality Management (TQM) Organization-wide commitment to quality at every level, emphasizing customer satisfaction, employee involvement, and preventive measures over inspection. PDCA (Plan-Do-Check-Act) cycle, benchmarking, employee empowerment Aerospace (Boeing's quality audits), consumer goods (consistent product standards) Builds long-term culture; improves overall satisfaction Slow to implement; vague without metrics
Theory of Constraints (TOC) Identifies and exploits bottlenecks (constraints) to maximize throughput; focuses on system-wide optimization rather than local efficiencies. Drum-Buffer-Rope (scheduling), five focusing steps (identify/exploit/subordinate/elevate/repeat) Metal fabrication (optimizing machine queues), supply chains (inventory buffers) Quick wins on throughput; holistic view Over-focus on single constraints; needs ongoing monitoring
Agile Manufacturing Emphasizes flexibility, rapid response to market changes, and customization; borrows from software Agile (sprints, adaptability). Modular production, cross-functional teams, quick prototyping Consumer electronics (adapting to trends like wearables), fashion (fast fashion cycles) Handles volatility; fosters innovation Less efficient for stable, high-volume ops; higher coordination overhead
Just-In-Time (JIT) Produces/delivers goods exactly when needed, minimizing inventory; often a Lean subset but can stand alone. Pull systems, supplier integration, minimal buffers Food processing (fresh ingredients on demand), auto parts (vendor-managed inventory) Lowers holding costs; improves cash flow Vulnerable to supply disruptions; requires reliable partners
World Class Manufacturing (WCM) Holistic framework integrating safety, quality, cost, delivery, and environment; pillar-based (e.g., focused improvement, autonomous maintenance). 10-20 pillars (e.g., TPM for maintenance), audits/scoring Fiat/Chrysler plants (pillar audits), heavy industry Comprehensive; measurable progress Complex rollout; resource-intensive
Sustainable Manufacturing Integrates environmental/social responsibility with efficiency; reduces resource use, emissions, and waste for long-term viability. Life-cycle assessment, circular economy principles, green metrics Textiles (recycled materials), energy (renewable sourcing) Eco-compliance; appeals to modern markets Higher upfront costs; regulatory variability

How to Use and Extend

  1. Start with Core Patterns: Check /patterns/core/isa95-hierarchy.yaml for a base UNS template. Customize by adding levels based on the types/methodologies above.
  2. Prototype: Use included C# or Python snippets to build a simple MQTT publisher/subscriber. Deploy to your home K8s lab or EKS for testing—focus on prototyping first, add ACLs/JWT later for security.
  3. Extend for Your Industry: Fork and add subdirs like /patterns/pharma/ with tailored YAML. For UI (if needed), suggest Vue3 with Vuetify for quick dashboards.
  4. Discuss: Open issues for ideas on variations, trade-offs, or integrations (e.g., bridging to RDBMS for queries).

This is just a starting point—let's iterate together.

Contributing

Pull requests are encouraged! Please follow a simple PR template: Describe the change, why it fits (e.g., aligns with ISA-95), and any backend examples. Keep it collaborative.

License

MIT License – feel free to use, modify, and share as permitted by this licensing model.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published