Automation Pyramid and the Unified Namespace | Unified Namespace (UNS) Basics 1

Summary

In this video, Dennis discusses the automation pyramid's role in organizing manufacturing data per the A95 standard, bridging operational technology (OT) and information technology (IT). He highlights four issues with the pyramid for data analytics: data aggregation loss, firewall barriers, integration hurdles, and transmission inefficiencies. He then explores how the unified namespace addresses these issues, with a conservative European approach suggesting its use for data analytics alongside the pyramid for process control. The video also touches on data flow organization and historical data storage options.

Takeaways

• 🔄 **Automation Pyramid Concept**: The automation pyramid is a data structure that organizes manufacturing data and systems, following the A95 standard.
• 🌐 **Bridge Between OT and IT**: It serves as a bridge between operational technology (OT) and information technology (IT), with OT at the lower levels and IT at the higher levels.
• 📊 **Data Aggregation Loss**: There is often a loss of data granularity as it moves up the pyramid due to aggregation.
• 🔒 **Firewall Barriers**: Security measures like firewalls create barriers that limit data access for analytics, especially at the lower levels.
• 🔄 **Integration Hurdles**: The multi-layered structure of the pyramid can create integration challenges when trying to access data across different levels.
• 🚀 **Transmission Inefficiencies**: There are inefficiencies in data transmission due to the difference between OLTP and OLAP databases.
• 🌿 **Unified Namespace Solution**: The unified namespace is proposed as a solution to address the limitations of the automation pyramid.
• 📈 **Two Views on UNS**: There are differing views on how the unified namespace should be implemented, with a more conservative European approach versus a more radical American approach.
• 🔌 **Connecting Systems**: The unified namespace connects to the automation pyramid to handle big data processing and analytics.
• 💾 **Historical Data Storage**: Historical data from the unified namespace can be stored locally in a historian or in a data warehouse or data lake in the cloud.
• 🔗 **Further Learning Resources**: Additional resources are provided for those interested in learning more about the unified namespace and its construction.

Q & A

What is the automation pyramid?

The automation pyramid is a data representation that organizes manufacturing data and systems in a plant, following the A95 standard, which is a common structure adopted by plants worldwide.

How does the automation pyramid bridge the gap between OT and IT?

The automation pyramid forms a bridge between Operational Technology (OT) and Information Technology (IT). Lower levels (zero and one) are associated with OT, while higher levels (three and four) are associated with IT.

What is the primary purpose of the automation pyramid?

The automation pyramid was designed for process control, not for data analytics.

What is the first problem associated with the automation pyramid?

The first problem is data and aggregation loss, where data produced at a high frequency by PLCs is often aggregated at a lower frequency before reaching upper systems, resulting in loss of intermittent information.

What are firewall barriers and why are they necessary?

Firewall barriers are necessary to secure production equipment, especially at level zero, where data can influence physical machine movement. They restrict access to those machines.

How does the structure of the automation pyramid with five separate layers affect integration?

The five separate layers of the automation pyramid create integration hurdles, as applications may need to connect to multiple layers to access data from different systems.

What are transmission inefficiencies in the context of the automation pyramid?

Transmission inefficiencies refer to the difference between OLTP (Online Transaction Processing) and OAP (Online Analytical Processing) databases. OLTP databases are designed for quick data writing but not for processing large data chunks required for analytics.

How does the unified namespace address the problems of the automation pyramid?

The unified namespace addresses these problems by either replacing the automation pyramid or, more conservatively, keeping process control with the automation pyramid and moving data analytics responsibilities to the unified namespace.

What is the difference between the American and European views on the unified namespace?

The American view proposes replacing the automation pyramid with the unified namespace, while the European view, especially at the UMH, is more conservative, keeping process control with the automation pyramid and moving data analytics to the unified namespace.

How is the data flow organized in reality between the unified namespace and the automation pyramid?

The data flow is organized by connecting both systems and reading all data from various levels into the unified namespace. There is also a data flow from the unified namespace towards the automation pyramid, which requires additional security measures.

Where is the historical data from the unified namespace typically saved?

Historical data from the unified namespace can be saved either in a historian locally in the plant or written to a data warehouse or data lake in the cloud.

Outlines

00:00

🤖 Automation Pyramid and Unified Namespace

Dennis, a Developer Advocate at the United Manufacturing Hub, introduces the automation pyramid, a data representation standard (a95) that organizes manufacturing data across plants. The pyramid bridges operational technology (OT) and information technology (IT), with OT-related data and equipment at the lower levels and IT at the higher levels. However, the pyramid faces four issues: data aggregation loss, firewall barriers, integration hurdles across five separate layers, and transmission inefficiencies between OLTP and OLAP databases. The unified namespace is proposed as a solution, with different views on its implementation—either replacing the pyramid or complementing it for data analytics, keeping process control with the pyramid and analytics with the unified namespace.

05:03

🌐 Unified Namespace Implementation and Data Management

The unified namespace is discussed in terms of its implementation and interaction with the automation pyramid. It contains a snapshot of the latest business data but does not replace the historical data storage function of the historian. The video script mentions two options for historical data storage: local historian systems within the plant or cloud-based data warehouses/lakes. Security measures are emphasized for data flow that could influence production processes. The video also references other resources for further understanding, including a video on the five UNS basics and articles linked in the description.

Keywords

💡Automation Pyramid

The Automation Pyramid is a conceptual framework used in manufacturing to organize data and systems within a plant. It adheres to the A95 standard, making it a universal structure for most plants globally. The pyramid is divided into five levels, with levels zero and one primarily associated with Operational Technology (OT), involving sensors, PLCs, and machine data. As one moves up to levels three and four, the focus shifts towards Information Technology (IT). The pyramid is designed for process control and not for data analytics, which is a key point in the video as it sets the stage for discussing its limitations and how they are addressed.

💡Unified Namespace

The Unified Namespace is presented as a solution to the challenges faced by the Automation Pyramid, particularly in the context of data analytics. It is a system that aims to provide a single source of truth for all business data, potentially replacing the Automation Pyramid according to some views. However, the video suggests a more conservative approach where the Unified Namespace complements the Automation Pyramid by handling big data processing and analytics, while the pyramid continues to manage process control.

💡Operational Technology (OT)

Operational Technology refers to the hardware and software used to monitor and control industrial plant processes. In the context of the Automation Pyramid, OT is associated with the lower levels (zero and one), where direct interaction with sensors, PLCs, and machine data occurs. OT is crucial for the physical operation of machinery and processes within a plant.

💡Information Technology (IT)

Information Technology, in contrast to OT, deals with the management of data, particularly for business insights and analytics. The video discusses how, as one moves up the Automation Pyramid to levels three and four, the systems start to operate more within the IT realm. However, the pyramid's design is not optimal for IT functions like data analytics, which is where the Unified Namespace comes into play.

💡Data Aggregation

Data aggregation is the process of collecting data from different sources and summarizing it. The video points out that the Automation Pyramid often results in data and aggregation loss, as data from PLCs, which may be produced at a high frequency, gets aggregated (e.g., by the minute or second) by the time it reaches upper systems. This aggregation leads to a loss of detailed, intermittent information, which is problematic for detailed data analytics.

💡Firewall Barriers

Firewall barriers are security measures that protect production equipment, particularly at the lower levels of the Automation Pyramid. They control access to machines and systems to prevent unauthorized interactions that could affect physical operations. The video explains that these barriers can limit data scientists' access to raw data, which is necessary for detailed analytics.

💡Integration Hurdles

Integration hurdles refer to the challenges of combining data from different layers of the Automation Pyramid. The video uses the example of building an application that requires data from an MES system (level two), a historian, and additional sensors, which would necessitate connections to three separate layers. This complexity increases development time and effort.

💡OLTP Databases

OLTP (Online Transaction Processing) databases are designed for handling transactions quickly, such as those found in level three and four systems of the Automation Pyramid. The video explains that these databases are not well-suited for processing large chunks of data, which is common in analytics. This limitation leads to transmission inefficiencies when trying to move large datasets through the pyramid for analysis.

💡OAP Databases

OAP (Online Analytical Processing) databases are optimized for analyzing large datasets, which is a requirement for data analytics. The video contrasts OLTP databases with OAP databases to highlight the inefficiencies in using the former for analytical tasks, which is a key reason why the Unified Namespace is proposed as a solution for handling big data analytics.

💡Data Flow

Data flow in the context of the video refers to the movement of data between the Automation Pyramid and the Unified Namespace. The video discusses how data is read from various levels of the pyramid into the Unified Namespace for analytics. It also mentions the reverse data flow, from the Unified Namespace back to the pyramid, which requires additional security measures due to the potential impact on the production process.

💡Historian

A historian in the context of the video is a system that collects and stores historical data from industrial processes. It is mentioned as a potential place to save historical data coming from the Unified Namespace, either locally within the plant or in a data warehouse or data lake in the cloud. This is important for long-term data retention and analysis beyond the immediate, snapshot data held in the Unified Namespace.

Highlights

Introduction to the concept of an automation pyramid

Automation pyramid organizes manufacturing data and systems following the A95 standard

Bridges the gap between Operational Technology (OT) and Information Technology (IT)

Data and equipment at lower levels belong to OT, while higher levels operate in the realm of IT

Automation pyramid designed for process control, not for data analytics

Problem 1: Data aggregation loss in the pyramid

Problem 2: Firewall barriers due to security concerns with production equipment

Problem 3: Integration hurdles due to separate layers in the pyramid

Problem 4: Transmission inefficiencies between OLTP and OLAP databases

Unified namespace as a solution to bridge the gap between process control and data analytics

Two views on unified namespace: American view vs. European view

Conservative European view suggests keeping process control and data analytics separate

Unified namespace connects to the automation pyramid and reads data from various levels

Data flow from unified namespace to automation pyramid requires additional security measures

Unified namespace contains a snapshot of the latest business data

Historical data storage options: local historian or cloud-based data warehouse/data lake

Clarification that unified namespace does not replace the automation pyramid

Link to a video explaining data flow organization in reality

Link to a video about the five UNS basics

Transcripts

[Music]

hi I'm Dennis developer Advocate at the

United manufacturing Hub in this very

short video I'm going to highlight the

concept of an automation pyramid and how

it relates to the unified Nam space so

the automation pyramid is really a data

representation that organizes your

manufacturing data and Data Systems of a

plant it follows the a95 standard which

means that nearly all plants in the

world will follow this similar structure

and the idea is that it forms a bridge

between OT or operational technology and

it where on the bottom levels levels

zero and one especially most of the data

and the equipment really belongs to the

realm of OT you have sensors n plcs

machine data and then the higher up the

you go towards level three and four

you're essentially operating in the

realm of it and in the middle is where

those two worlds essentially come

together and it's important to note that

automation pyramid was designed for

process control it was not designed for

data analytics and we will see why in

the following four problems first of all

you always have a data and aggregation

loss in the pyramid this means means

that for instance your PLC commonly

produces data at 20 millisecond

frequency but by the time this data

reaches your upper systems it would have

been aggregated for example by minute or

by second in the historian or even worse

by batch number in the MS system so you

lose all the intermittent information

because it has been averaged the second

problem is that you have firewall

barriers these are necessary because

you're dealing with production equipment

especially at level zero so your data

can actually influence how machines move

in the physical world which is why you

need a lot of security about who can

access those machines and who can't well

in practice for data analytics this

means that your data scientist can only

get access to b c's Often by actually

going through the physical process and

connecting his laptop there instead of

jumping through a lot of um Networks

so a third problem that we have is that

the fact that automation pyramid

consists of five separate layers is that

you have integration hurdles imagine

you're building an application that

requires data from the mes system but

also some data from the historian so

from level two and maybe some additional

sensor well this means that your app

will have to connect three times to your

automation pyramid to each of those

separate layers which greatly increases

the the time of

development and finally you have

transmission

inefficiencies and that's because we

have the difference between oltp

databases and oap databases where the

former are more meant for transactions

for example your level four and level

three systems are meant to quickly write

data but they are not designed to

process large chunks of data as often

happens in analytics where you calculate

for instance an average over month or

you process half a year of data

so even if you would be able to Route

all your data upwards to towards from

Level zero to level four your level

three and level four systems will not be

able to handle that large data flow so

how does the unified namespace um

address this problem well we have to

distinguish between two views the more

American view proposed by wer Reynolds

says that in the end we're going to

replace the automation pyramid by the

unified name space which becomes your

ultimate source of truth and the main

system on which your entire business

runs now in the in Europe especially at

the umh we are a more conservative view

in the sense that we will want to keep

separating process control and data

analytics where process control will be

done by the automation pyramid and all

the responsibilities of big dat

processing and analytics we will move to

the unified

namespace we do this by

connecting both systems and by reading

in all the data from the various levels

into the unified name space if you want

to know more about how this data flow is

organized in reality I would recommend

to watch the video I linked below where

I explain which data lands on the mqtt

Brokers there's also a data flow in the

opposite direction so from the Unified

namespace towards the automation pyramid

now this data flow is more risky in the

sense that it could potentially

influence the production process which

is why additional security measures must

be

implemented finally the unified

namespace in essence only contains a

snapshot of your business with the

latest data so the question arises where

do we save the historical data coming in

unified namespace and for that we will

either save this data in the historian

locally in the plant this is what the

umh also offers or we can choose to

write this data to a data warehouse or

data Lake somewhere in the cloud so I

hope this clarifies the automation

pyramid and how it interacts with the

unified namespace in a sense that we

will not replace uh that the unified

doesn't replace the automation pyramid

if you want to know more about how the

unified Nam space is um build I have we

have a video about the five UNS Basics

here linked above where we go through

each component of the unified name space

as always I will link the two discussed

articles Below in the description and I

will see you in the next video bye-bye