UNS – The Sole Key to Industrial Digitalization
I4.0 Insight
3 minutes
Oct 10, 2024
It's not entirely wrong to suggest that every industrial digitalization project to date has fallen short. A successful business model in this sector requires three key factors: achieving a measurable positive ROI, standardizing for scalable replication, and replacing holistic systems with modular apps. In reality, no industrial digitalization project on the planet has yet met all three of these criteria.
The technological evolution in the industrial sector has long been a process of waiting for breakthroughs. However, since the advent of Industry 4.0, the manufacturing industry has grown increasingly anxious, coining various terms to prove that digital transformation has already become a mature and fully successful market.
We are paying the price for this, most notably seen in the disappointed expressions of engineers on both the user and supplier sides when discussing digitalization. The paths we've attempted—OPC UA, cloud-native solutions, OLAP, and low-code—have all proven unworkable in industrial environments, failing to meet the 3 essential elements for success.
Yet now, a new buzzword, "UNS/Unified Namespace," is quietly gaining traction. This time feels different—it might just be the key to achieving the elusive "impossible triangle" we've long dreamed of.
One UNS Platform
Jeremy from UMH was the first to provide a detailed definition of UNS technical architecture. He created a chart that reveals the best practices for implementing UNS.
from UMH Blog- Integrating the Unified Namespace into Your Enterprise Architecture
In essence, UNS functions as a data pipeline that converts various protocols into MQTT messages, which are then sent to an MQTT broker for real-time consumption by various applications, while being stored in modern databases for historical analysis. Applications built on the UNS architecture are naturally event-driven, eliminating the need for repeated polling of unchanged tag values in the design. The data flowing through the system inherently includes key event information, such as shutdowns, startups, and orders, allowing developers to focus on less while achieving more.
Specifically, due to the fragmentation of technical components, the entire UNS stack contains many intricate details. A significant portion of this complexity is aimed at ensuring the consistency and integrity of data transmission. Kafka is paired with MQTT to compensate for MQTT's limitations in data persistence and issue tracking. Additionally, NodeRED's limited scalability has led UMH to adopt Benthos for converting certain standard industrial protocols.
However, UMH's southbound, northbound, and internal data interface formats and processing methods have already become highly standardized. It’s foreseeable that, regardless of how UNS evolves, the integration and storage layers based on MQTT and modern time-series database technology are now fixed. This represents a level of standardization we’ve not seen in other industrial digitalization tech stacks.
Two Automation
Modern apps, particularly web apps, are typically composed of a backend (data and services) and a frontend (HMI, UI/UX).
The backend needs of industrial apps can generally be divided into two categories:
Business Automation
This includes applications such as batch records, shift handovers, and inspections. These apps process and manage low-frequency relational data to ensure that events and transactions follow the designed workflows. In these cases, humans are the primary operators.
Process Automation
This includes applications like SCADA, APC, and RTO, which analyze high-frequency time-series data and use time-series processing to optimize industrial processes. Here, the operators are actuators and programs, not people.
These two backend automation needs can be simultaneously addressed by an event-driven architecture based on UNS. Since industrial applications typically do not require billions of TPS (transactions per second) or handle high-concurrency scenarios like flash sales, an MQTT-centered data distribution system can effectively manage both relational and time-series data requirements.
Conversely, front-end requirements cannot be standardized or even categorized. With the rise of web technologies, the flexibility of industrial human-machine interfaces (HMI) and human-computer interactions (HCI) has surged. Complex icons, historical trend analysis, dynamic rendering, Web3D, and XR are now all within reach. This is also evident from the frontend technology ecosystem, which undergoes monthly iterations.
From the front-end perspective, industrial apps should thrive in diversity and maintain the utmost flexibility. Or there should be industrial apps to adapt swiftly to the ever-evolving technology trends and user requirements. Implementing this is straightforward: leveraging IT technologies like React, Vue, and Node.js is sufficient. However, the choice of specific technologies should be left to the developers, as any rigid definition of the front end imposes unnecessary limitations.
Three components to success = 1 + 2 + N
Let's move back to the foundational elements required for successful industrial digitalization: ROI, standardization, and Modular Apps.
The unclear ROI of digitalization projects stems from the still nebulous scope and boundaries of industrial digitalization. Often, businesses hesitate to or are unable to tackle tough challenges such as control optimization, supply chain forecasting, or replacing traditional MES/ERP systems. Instead, they opt for solutions like paperless operations, instant messaging, or BI dashboards—which are not critical needs at all.
Digitalization should encompass both production management and business operations, penetrating every aspect of these areas and even disregarding the boundaries defined by previous technology.
Standardization involves identifying the greatest common denominator within the technology stack and establishing consensus. Within the realm of IIoT, this common denominator is the Unified Namespace (UNS). MQTT offers a universal model and addressing method for industrial protocols without being overly restrictive. This approach prevents the complexities and maintenance challenges often associated with systems like OPC UA and its related standards.
from i-flow
We can locate relevant standards and functions for all our needs within the OPC UA manual; however, the downside is that this manual has exceeded a thousand pages. You can imagine the difficulty in replicating and sharing this vast document with another engineer.
A major challenge of Modular Apps comes from the digitalization suppliers themselves. In our pursuit to simplify app development, we attempted to invent a low-code silver bullet for developing industrial applications, which ironically generated complexity and limitations.
For instance, Thingworx introduced the Thing Centric concept, Meshup components, and low-code development environment, fantasizing that developers would one day embrace this “development platform”. However, they overlooked a simple fact: JAVA + JavaScript is already a form of low-code development compared to C/C++. It’s unrealistic to create a low-code platform that would make developers abandon their existing programming frameworks; otherwise, they wouldn’t be called developers but Thingworx engineers instead.
A comment from the official PTC forum punctures this supplier illusion:
Good technology should broaden the creative horizons of application developers.
A platform based on the UNS architecture, combined with two automation that span the industrial enterprise lifecycle, and add unrestricted apps, constitutes a successful model for digitalization.
The Fourth element of success
The network effect is the underlying logic of the internet era. Open source is the easiest path to achieving this effect in the software domain. We see components like NodeRED, TDengine/TimeScale, and EMQX/HIVE, which align with the UNS architecture, have already formed vast engineering and developer communities through open source. The combined reach of these communities can surpass decades of customer accumulation by commercial software companies. Within these community networks, the technology components are vendor-neutral, giving both engineers and users confidence to invest. This is the point that giants like Mindsphere, Predix, and Thingworx have long sought but failed to achieve.
If you are a user, we encourage you to adopt the open-source architecture of UNS and the application development framework of 1+2+N within your organization. This approach can help streamline operations, reduce costs, and facilitate the achievement of your objectives.
If you are an engineer, we invite you to contribute to UNS and other industrial open-source communities. Your insights and critiques are invaluable as we strive to foster a more open industry. Join us in transforming the landscape of industrial technology!
