Factories can do a lot better by digitizing collaboration

November 1st 2019 by Maximilian Fischer

Digital collaboration on the shop-floor offers the ability to fundamentally change how factories operate today. Read on to learn more about what this means and how Actyx helps factories to capture this opportunity.

Collaboration in the factory context

Let's first look at what collaboration means in the context of a factory. Collaboration is, in essence, the action of working with someone to produce something. That means translated to a factory that humans, machines, and IT-systems are collaborating to produce the right product at the right time and in the right quality.

A factory is a complex system, in which often many dozens or hundreds of stakeholders are collaborating. The production planner needs to collaborate with supervisors and machines to understand resource availability, order progress and scheduled maintenance to create a production plan and communicate it to the operators. A logistics worker collaborates with the planner, production operators and other logistics workers to understand, which material she should deliver to which workstation at what time.

Setup personnel collaborates with machines to correctly adjust the machine for the right product. Machines are also collaborating directly, e.g. a robot with a CNC-machine to load and unload material. IT-systems can also collaborate with machines, for example when a new recipe is automatically loaded from the ERP system to a production line. These forms of collaboration are depicted in the following graphic.


Collaboration in a factory can be compared to an orchestra, in which many musicians playing different instruments need to work together to perform an act. Each musician needs to play the right notes at the right time reacting to signals from the conductor.

Note: There is, of course, a lot more collaboration happening in a factory between engineering, sales, purchasing, HR and more. Throughout this article only collaboration on the shop-floor is considered, though.

Factories can do a lot better

Factories around the world produce an incredible amount of high quality products, providing the basis for our wealth and standard of living. However, factories are still creating a lot of waste when manufacturing these products. Ineffective collaboration is a huge factor that reduces efficiency, quality, and flexibility in today's factories. This results in a huge amount of waste: it is not extraordinary for factories to spent 95% of all work time performing non-value adding work, i.e. work that adds no value to the finished product (Source).

Note: factories will never spend 100% of their resources on value-adding work, as non-value adding work also includes things such as setup of machines, inspection work or transportation of material. These tasks cannot entirely be eliminated.

There are three main factors why collaboration is not as effective as it could be in today's factories:

  1. Collaboration involves communication overhead
  2. Information propagation and decision making is slow
  3. Decision-making is often wrong or inaccurate due to inaccessible information or data

As a simplified example consider a factory consisting of two workstations, in which workstation 1 is producing a product and workstation 2 is controlling the quality of this product. The two workstations are spatially separated and material is manually transported from station 1 to station 2. At both stations, the work is done by machines (machine 1 and machine 2). Machine 2 at workstation 2 now detects that the product measurements are out of tolerance. The product is still being produced at workstation 1 by machine 1. In this scenario, the machine at workstation 1 needs to be adjusted to compensate for the quality error at station 2. The setup is illustrated in the following graphic.


Note: there are many reasons for a quality defect, and adjusting machine parameters is only one possible cure. We assume for this example that it is possible to compensate for the quality defect by adjusting the machine settings.

In the ideal world, machine 2 collaborates directly with machine 1. Machine 1 is informed of the precise type of the quality defect by machine 2, can access sensor data on input material properties, the health of its mechanical components, and has access to historical sensor and quality data. Based on this information, machine 1 decides to adjust its settings to compensate for the defect. Machine 2 provides direct feedback on the products that were produced with the new settings. This is illustrated in the graphic below.

Ideal reaction to quality defect

Though, this is not how it usually works in factories. Typically, an operator would detect that the machine reports a quality defect. The operator would then inform their supervisor, who in turn would inform the quality assurance responsible and the supervisor of machine 1. After analysing the problem and based on their experience, they would decide how to adjust the machine. Somebody would inform the operator, who then would make the adjustments. In order to fix this problem, at least 5 stakeholders would need to collaborate. On top of that, the decision on how to fix the problem is based on gut feeling more than it is on data analysis, and it is a lot slower than if the machines would collaborate directly.

Actual reaction to quality defect

This is just one simplified example of why collaboration is inefficient in factories. Picking up the analogy of an orchestra, if the factory was an orchestra, it would consist of 20 musicians playing instruments (value-add) while 80 people are assisting them (non-value add).

This inefficient collaboration will become an even greater problem for factories in the future. Customers demand more and more complex and individualized products. Factories need to become more flexible whilst maintaining or even increasing the level of quality. This will not be possible for most factories unless they are significantly changing how they operate today.

The huge opportunity for factories

But not all is lost for factories, in fact, quite the opposite. There is a huge opportunity for factories to outpace the competition by digitizing collaboration and increase efficiency, flexibility, and quality. Digitizing collaboration offers several key advantages:

  • collaboration is more direct, i.e. it occurs only between relevant stakeholders
  • collaboration can occur over a much larger set of stakeholders without compromising on speed
  • decision-making can be automated to make it a lot faster
  • decision-making can be based on data and becomes more accurate

It offers the chance to reduce non-value adding work to a minimum. By digitizing collaboration, the factory becomes a system of interconnected stakeholders, which are constantly communicating to coordinate execution of work, find solutions to fix problems, optimize processes and automatically share learnings. Only through this digital collaboration is it possible for factories to truly optimize processes across the shop-floor, eliminate bottlenecks and become more efficient and flexible.

View of a factory transitioning from analog to digital

For concrete examples of digital collaboration scroll down to the bottom of this article to the section Appendix: Examples of digital collaboration.

How to digitize collaboration

A stakeholder that wants to collaborate digitally needs two things: (1) the ability to process information and make decisions, and (2) the ability to communicate with peers to exchange information.

A simple example of a digital collaboration of a machine and a human is depicted in the graphic below. In this example, a machine collaborates with a human to initiate a reaction to a machine interruption. For this, the machine needs to be able to process information from its sensors and be able to detect that it is broken. The machine needs to be able to communicate this information to the operator and the operator needs to process this information to understand that the machine is interrupted.

What is needed to digitise collaboration

Translated to the world of IT this means that each stakeholder, that wants to collaborate, needs:

  1. an interface to the stakeholder, i.e. the machine, human or IT system
  2. the ability to run logic to process data and take decisions
  3. the ability to communicate data in real-time between peers

Realizing digital collaboration across many different stakeholders is extremely difficult and costly today. We at Actyx believe that digitizing collaboration needs to become a lot simpler to truly allow factories to reach the next level of efficiency and flexibility. That's why we created ActyxOS, a developer platform that follows a radically different approach compared to traditional IT systems, making it significantly easier to digitize collaboration for factories.

Digitizing collaboration with Actyx

ActyxOS is a platform for developers to rapidly implement digital collaboration use-cases. It provides:

  • a runtime to execute logic on edge-devices
  • connectivity between edge-devices in a peer-to-peer network
  • persistent data storage across a network of edge-devices

On top of ActyxOS, we offer a programming model — Actyx Pond — that simplifies the development of Apps on our decentralized platform. With these functionalities, our platform optimally fulfills the requirements of digitizing collaboration discussed above. Our platform offers the necessary tools and infrastructure allowing developers to fully concentrate on developing value-creating business logic.

Access our documentation to learn more about ActyxOS on the Actyx Pond.

In contrast to most vendors, our platform is fully decentralized. Traditionally, collaboration is digitized with a centralized architecture. Client hardware devices such as desktop computers, PLCs or mobile devices are connected to a central node, which is responsible for running logic, storing data and communicating to and from client devices. This approach holds two main problems, especially in the context of digitizing collaboration in factories:

  • Complexity in connecting peers: connectivity occurs point-to-point, in a defined hierarchy. This becomes ever more complex the more devices need to be connected. Factories require dozens if not hundreds or thousands of devices to be connected, which becomes unmanageable at some point with a centralized approach.
  • Reliability in running software: the central nodes are critical for the operability of the systems. If the central nodes are down or can’t be reached by the devices the entire system doesn’t work anymore. This could mean that operators, for example, don’t have access to information on how to set up a machine, causing production interruptions.

In contrast to this centralized approach, our platform requires no central server and logic is executed entirely on edge-devices. Edge-devices are communicating directly in a peer-to-peer network. Connectivity becomes easier as no complex hierarchies need to be built; single-points-of-failures are eliminated, making the whole system more resilient. The difference between these two systems are depicted in the graphic below:

Central vs. decentralized architecture

To come back to our example of the collaboration of a machine and an operator in case of an interruption. This collaboration could be realized by running an app on a gateway that interfaces with a PLC in the control cabinet of the machine. The app analyses the sensor data coming from the machine and detects an interruption based on pre-defined conditions. It directly communicates with a tablet that the operator carries with him. An app on the tablet processes the incoming information from the machine and visualizes the interruption for the operator.

Digitising collaboration with Actyx

With our platform, we make it significantly cheaper and faster for factories to implement digital collaboration use-cases when compared to a traditional, centralized approach. This opens up the opportunity for many factories to digitize processes that were too costly to digitize with a traditional approach. It also paves the way for the rapid implementation of innovative use-cases, in which many stakeholders collaborate or in which machine-to-machine collaboration plays an important role.

We’ve illustrated a few example use-cases that can be realized on the platform in the Appendix of this article. Inspired by these, you’ll probably think of other cases as well — whatever it is, we are looking forward to seeing your ideas come to fruition!

Contact us if you want to talk about how your use-case can be implemented on our platform or if you want to get hands-on experience building your first app.

Appendix: Examples of digital collaboration {#example-use-cases}

This section contains concrete examples of digital collaboration use-cases. These use-cases are only a small set of all possible use-cases that are suited to be implemented on ActyxOS.

Use-case 1: Improved quality through automatic detection of quality deviations and adjustment of machine parameters

Machines are directly collaborating to quickly compensate for a quality defect. A machine detects a quality problem and automatically informs the relevant machine upstreams in the value chain. The upstreams machine analyses actual settings and sensor data, compares it with historical data and automatically adjusts its settings to compensate for the quality defect.

Example use-case: quality assurance

Use-case 2: Quicker setups through the assistance of the operator with real-time sensor and machine data

An operator is digitally collaborating with a machine to effectively set up the machine. The machine analyses its real-time sensor data to inform the operator about the wear of its components. The operator can access this data on a mobile device and compare it with historical data, cross reference with settings and quality, and identify how to optimally set up the machine. The operator loads a new machine program via the mobile app and changes settings if necessary. Through this collaboration, the operator has all the relevant information to optimally and quickly set up the machine.

Exemplary use-case: setup assistance

Learn more about how Actyx digitized the setup processes for one of the leading European packaging glass producers on our success-stories.

Use-case 3: Optimized material flow through intelligent coordination and execution of intra-logistics material transports

Delivering material too early or too late to workstations either causes production interruptions or unnecessary stock. To optimally deliver material, logistics and production stakeholders need to collaborate. Autonomous guided vehicles, drones, and forklift drivers need to understand how much material is at each workstation, when the next order will be started and where to find the relevant material. They also need to collaborate with each other to coordinate who is picking up which material depending on their current availability, distance to the target location or current status such as battery level.

Exemplary use-case: digital intralogistics

Use-case 4: Optimal resource utilization through intelligent machine data analysis and digital allocation of work

Allocating work to resources is a complex task. The production planner needs to collaborate with all relevant resources to optimally allocate work. The planner needs to take into account many things such as when the current orders will be finished, which products can be produced on which resource, what is the best sequence of products on each resource to reduce setup time or when the machines need to be maintained.

In the digital factory, each machine analyses real-time data to compute when it needs maintenance and when it will be finished with the current order queue based on current settings. This data is then combined to compute a suggestion for the planner on how to allocate work to resources. The planner amends the plan and communicates new work orders directly to each workstation. The planner and the machines need to constantly collaborate in case of a machine interruption, slower production speed or quality problems. Only through digital collaboration and intelligent analysis of data is it possible to effectively allocate work to resources.

Exemplary use-case: resource allocation

Use-case 5: Modular machine configuration through automatic reconfiguration of production lines

Products are becoming more and more individualized, requiring constant change of machine configurations. Static production lines will be replaced by modular machines that can be flexibly combined. This requires machines to directly collaborate to configure themselves to production lines. Machine modules are automatically detecting each other in case they are physically connected, are exchanging configuration parameters and are automatically adjusting their settings to produce the relevant product.

Exemplary use-case: flexible machine configuration