What is an Automation Pyramid?

The Automation Pyramid structures automated production entities into layers of specified tasks and corresponding IT infrastructure. The general level set separates a production site traditionally into six levels, forming a pyramid of high complexity at the bottom and sharp focus at the top. Data at lower levels is often time critical and needs to be processed directly at the edge of an IIoT installation.

In recent approaches to industry 4.0 nodes at the edge allow to flatten the Automation Pyramid by combining level 1 and level 2 (sometimes also level 0) into cyber-physical systems or smart production entities:

Automation pyramid with six levels from the process level with sensors and actuators to the enterprise level with ERP

Process level – level 0

Production Process:

Simple sensors and actuators provide data and execute commands (e.g. sample temperature and programmatically turn on power for a fan).

Field level – level 1

Sensing and Manipulating:

Collects data and distributes commands, often using field bus and similar infrastructure (e.g. group multiple temperature data and instruct a device throttle workload).

Control level – level 2

Evaluating and Controlling:

Programmatically controls the production process by evaluating sensor data and issuing commands (e.g. analyze temperature change for rising temperature and assess counter actions to prevent damage). Standards such as OPC UA describe the data of this level so that higher levels can use it.

Operations level – level 3

Monitoring and Supervising:

Offers a broader view through collections of datasets and at production devices. Also allowing recipe adaptions via process management systems (e.g. detect potential overheating and slow down production process for a device).

MES level – level 4

Manufacturing Operations Management:

Manages production sites through production planning, performance monitoring and quality management (e.g. inform about reduced production capacity due to temperature and reduce raw material demand).

Enterprise level – level 5

Business Planning and Logistics:

Provides general data for production planning like capacity limitations and connects order management systems and other relevant information from ERP systems (e.g. increase delivery times, prioritize fixed date orders, divert production to other plants, management may decide to install an air condition, …).

Automation Pyramid and Unified Namespace: Why Data No Longer Travels Level by Level

In the classic pyramid, systems are connected level by level, and every additional connection across levels is yet another interface of its own. A temperature reading from the field level only reaches the MES once the control and operations levels have passed it on. Every new system that needs machine data requires another integration project. The more applications need access to production data, the more this point-to-point structure slows things down.

A Unified Namespace (UNS) turns this principle around: all levels publish their data to a central, consistently structured data layer, and every system accesses it there directly, without going through the hierarchy. The levels of the pyramid don’t disappear, they still describe which system takes on which task. But the data flow no longer follows their structure. How this works in practice and how to build a UNS is explained in detail in our glossary article on the Unified Namespace.

Comparison: automation pyramid with point-to-point interfaces and Unified Namespace with a central hub

Connecting Every Level of the Pyramid with Cybus Connectware

The levels of the automation pyramid remain in every factory: PLCs, SCADA, MES and ERP each have their own tasks and their own interfaces. Cybus Connectware connects these systems through a shared data layer instead of connecting each level individually to each application. Machines and controllers are connected once, and their data is then available in structured form and in real time for MES, cloud platforms and AI applications. At the same time, you define which system and which person may access which data.

This lets you add new use cases and additional plants without rebuilding the integration for every project.