Manufacturing the Data Mining on Edge


Verena Stanzl – Data Scientist, Fill GmbH

Latest achievements show that edge computing is an emerging technology field. By 2022, 90% of industrial enterprises will utilize edge computing (Frost&Sullivan,

This is not surprising - considering the many advantages that Edge Computing offers:

  • Speed : Relevant data needs to be processed fast. Edge Computing reduces the overall volume of data traffic increasing the performance of applications and services. On the other hand it enables latency-sensitive applications, where decisions have to be made fast to prevent failures or accidents. This would not be possible if data has to be transferred to a data processing center.
  • Security : The number of IoT devices is constantly growing making them a target for potential network attacks. Edge devices can filter sensitive information and, if necessary, transmit only non-critical data to meet security and compliance requirements.
  • Costs : With edge computing, data can be filtered at the point of origin and does not need to be sent to a data center entirely. Data processing and storage in edge devices reduces expensive bandwidth requirements and thus optimizes overall costs as using a mix of local services and cloud-based applications provides a cost-effective IoT solution.
  • Scalability : As business grows, more data and more computational power is necessary. With Edge Computing, a less expensive alternative to a dedicated data center is possible by combining IoT devices and edge data centers. It also reduces growth costs, as adding more devices don’t greatly increase the network’s bandwidth demands if preprocessing and filtering is performed on the edge device.

Edge Computing in Manufacturing

The digitization of manufacturing machines and the use of the data produced in the process shows great potential for process optimization to increase profitability and product quality whilst enabling new business opportunities and strategies.

The advantages of edge computing can be fully exploited. Processing real-time data can be a challenge – but with Edge Computing data processing near its origin enables faster decisions. For example, the identification of unstable processes and appropriate intervention requires low-latency to prevent accidents and failures.

Process data contain sensitive information about the underlying processes and expert knowledge. Therefore, some data should not be transfered to the cloud to meet security compliances – staying on the edge.

If all the machines in a factory are connected, the amount of data and computing power required can skyrocket, considering a machine produces several GB of data per day. By pre-processing and filtering the data on edge, network bandwidth requirements and cloud costs are maintained, as only data necessary for further analysis is transferred to the cloud.

Cybernetics is FILL’s Industrial IoT solution. Running analysis on edge it enables customers to increase their profitability and product quality by monitoring the condition of the machines during production. By identifying deviations, they can intervene in time to ensure process stability and consistent product quality on high-standard.

Figure 1: FILL's Cybernetics enables monitoring of the machine's condition uitilizing Edge and Cloud Computing

As long as the calculations are only related to one machine, the capacities on edge are sufficient. But as soon as the analyses have to be performed across machines, or larger data sets and more computing power are required for machine learning methods, data transfer to the cloud becomes necessary.

Get the best of both Edge and Cloud Computing world

Combining edge and cloud computing enables the benefits of both to be combined, where latency-sensitive applications and services can run on the edge and applications with high computing power requirements will run in the cloud.

During PLEDGER the benefits of distributing data and computational power across edge and cloud can be investigated. FILL is running digital services on the edge and cloud and will exploit and evaluate the tools provided by the PLEDGER platform. In the course of this, the stability and performance effectiveness of the infrastructure used should be enhanced and the computational nature of the provided services will be understand investigating Quality-of-Services metrics.
Benchmarking the applications and deciding whether the service should run on the edge or in the cloud based on the required ressources will enable the optimization of FILL’s digital products and further development of Cybernetics. Using trust and smart contracts, new business models can be explored and competiveness will be extended.

The FILL Future Zone – FILL’s Center for Digitalization and R&D – is the perfect environment to integrate the PLEDGER platform and is ready for set-up and integration during 2021.


Figure 2: The standard machine tool syncromill is set-up in the Future Zone and ready for PLEDGER integration.


is loading the page...