RTA's Blog
Why Local Data Context Is Now Mandatory
Apr
In Industry 3.0, most data stayed inside the plant. SCADA could handle a lot of missing source data context because humans filled in the gaps. That does not work anymore. The transition to Industry 4.0 has eliminated the safety net of human intuition that once bridged data gaps, replacing linear data paths with a mandatory requirement for machine-ready context at the source.
What Is Data Context?
Data context is what turns a tag dump (a large collection of unsorted data) into something you can query, compare and act on. Data contextualization means taking raw measurement values and attaching the information that makes them usable. This additional information includes:
- The time the data was collected
- The identity of the device or process that generated it
- The physical asset and location
- The units of measure and the actual meaning (semantics) of the value
- The relationships that explain what the value represents
No context means no trust and no consistent interpretation. You cannot compare across machines or time, and you cannot automate decisions. AI is not the first thing that breaks. Basic SCADA, ERP integration and business intelligence all become fragile or meaningless.
How Can Data Context Impact Operational Systems?
One of the most famous examples of the failure to specify the correct data context was the disintegration of a $125 million spacecraft. Lockheed Martin provided thruster calibration data in pound‑force seconds. NASA’s Jet Propulsion Laboratory (JPL) assumed the data was in newton‑seconds. Because of missing context, the spacecraft’s trajectory calculations were off by a factor of 4.45 and the spacecraft disintegrated.
A similar, but much less significant, type of failure occurs frequently with temperature measurements on the plant floor. If it’s not stated, upper-level systems have no way of knowing whether a process temperature measurement is Centigrade or Fahrenheit. In most of these cases, there is prior information that provides the data context. That’s an important point: context always exists, but it’s not always specified.
How Has Data Context Changed with Industry 4.0?
In Industry 4.0, the world has gotten less linear. No longer does data move along a linear path from PLC to SCADA to MES and then to ERP, with additional context added at each step. Today, many plants have adopted a less linear architecture sometimes based on the Unified Namespace (UNS). Data flows directly from control systems into MES, ERP, cloud analytics and AI systems. This exchange can’t happen if the data isn’t given context at the source. MES and ERP systems cannot guess what your tags mean. They assume the data is already trustworthy and complete.
Where Should Data Context Be Added in a Control System?
Data context is best added at the data collection point, where there is the most knowledge about the location, process state and the data source. The farther upstream data context is added, the less reliable it is. Adding process context (Run state vs. Jog), Time Context, Location context and other data context upstream in an ERP or MES system is a questionable strategy, at best.
This data must typically be configured by a control engineer who has the direct knowledge to effectively add context to relevant data points. This is far more practical than having a data scientist guess what a data point is or how it interacts with other points in the data.
Why Don’t Edge Gateways Add Context?
Most gateways treat data like a delivery problem, not a meaning problem. They move tags. They convert protocols. They publish payloads. But they do not model context.
In practice, edge products fall into three groups:
- Basic protocol converters (most of the market) – They move data and maybe add a tag name. Context is usually entirely missing in these devices.
- Edge compute frameworks – They allow you to script or build models, but only if you invest serious engineering time. Powerful, but heavy and complex.
- Industrial data hubs – They support rich context models but are expensive and overkill for many cell-level use cases.
| Basic Protocol Converter | Edge Computer | Industrial Data Hub | |
|---|---|---|---|
| Capability | Add Tag Name | Some Modeling Capabilities | Rich Context Modeling |
| Engineering Resources Required | Minimal | Some Required | Significant |
| Limitation | Not typically capable of adding context | Meta data capable but little modeling | Extensive modeling capabilities but farther from the process |
| Investment | Minimal | Minimal | Significant |
The gap is at the machine and cell level, where most useful time-series data is generated but rarely contextualized correctly.
How to Add Data Context in a Control System
Since context is best added locally, often the right place to set context is the PLC rack, where process meaning still exists. A lightweight PLC historian with modeling capability can capture time-series data, apply identity, semantics, and process context, and then publish usable data upstream.
Don’t Just Collect Data, Make It Useable
Data becomes valuable when you collect it and add context to it so it can be clearly understood. Context is not a feature. It is the foundation. If your digital systems are not built around a strong data context, you do not have an Industry 4.0 architecture. You have a very expensive noise generator that no AI system can fix.
Act now to avoid costly mistakes and secure lasting value. Edge data collection solutions, like the Real Time Automation’s RTConnect A-B PLC Historian, let you add the right level of data context at every stage. Your information infrastructure—and your business future—depends on the clarity and integrity of your data. Take the next step today.
Have questions or need more information? Visit the RTA Learning Center or contact an RTA Enginerd application engineer at 262-436-9299 or solutions@rtautomation.com.
