5 June 2019

Edge Computing and Mindsphere – A True Dreamteam (4/4)

MindSphere and Industrial Edge will define the new structure of data processing for the factories. In this context, edge computing is used to optimize our cloud computing system by performing data processing at the edge of the network, which means, near to the source of the data.

(Read my previous blogs for a general introduction how we deal with it in our own factories: Connecting 30 factories, North Star principle Manufacturing Data Platform.)

Figure 3: Real time model execution on the edge and model development in the cloud

A key challenge is, how can we get our data processed from so many different devices. 

Edge computing – the new border to the web?

Edge computing with Industrial Edge is the option, but in difference to the manufacturing data platform architecture, that centralizes processing and storage in a single and massive data center, Industrial Edge pushes the data processing power to the edge devices. The main advantage is, that only the results of the data processing need to be transported over networks. Relating to the use case this provides precise and real-time results and consumes far less network bandwidth. With other words, the processing of the data is close to the source of the data.

Of course, this methodology will help us to significantly reduce the storage quantity. Often this data ends up being useless anyway and the fear exists that the company will waste thousands of dollars storing data that they almost certainly will never use. 

Industrial Edge and MindSphere complement each other

It makes absolute sense to divide the processing between the Industrial Edge and the centralized manufacturing data platform (MDP). The goal is to process the data that is needed quickly close to the device, in cases that need immediate action and where response time is essential, e.g. quality selection or corrective actions in machine applications. In such applications, there is no way that data could be send to the centralized MDP, processed and the response sent back in less than 100ms.

But one important thing has be considered with edge computing and looked at more closely. Edge computing does not store data in long term, it eventually gets deleted, which isn’t conducive to big data analytics and business models as DaaS (Data-as-a-Service) or AaaS (Analytics-as-a-Service). 

Figure 4: The way from a technical infrastructure to DaaS
and AaaS

As stated above, our edge devices only process locally collected data and in most cases this data is discarded after use. Thus, if you think you should store all collected data for cumulative analytics decision-making purposes like AI, then edge computing alone is not the right fit.This is where the MindSphere cloud based MDP has its huge value aside from processing data that is either not as time-sensitive or not needed by the device directly, such as for big data analytics on data from different devices. 

We are going to create purpose-built Industrial Edge computing-based applications in each plant in combination with the MDP more than 30 plants will be connected to form a fully integrated data network.

On the edge layer, for example, will be applications that place data processing in sensors to quickly process reactions to alarms like our spindle use case, where we predict maintenance for de-panelling machines, or the which decides if a printed circuit board must be tested or not in the range of milliseconds. 

Figure 5: Architecture of the x-ray use case 

On the other hand, we are not going to place our inventory-control data and applications at the edge or time-relevant content drift observations to train and strengthen smart algorithms — this would result in a distributed, unsecured, and unmanageable mess with no option of scalability of solutions. This is why the training of an algorithm to become stronger will be done in the cloud and the insights may serve later for an AaaS business.   

In this context, edge computing will not replace cloud computing via a manufacturing data lake nor vice versa, but the two approaches will complement each other. To state that edge computing will displace cloud computing is like saying a PC would displace a data center.

It will provide data scientists and analysts with the possibility to proof hypothesis and look for correlations and patterns, as well as enabling business and service user, shop floor engineers and technicians to explore data, define shop-floor relevant use cases and create reports by their own. 

This approach of combining edge with cloud computing will allow scalability of IoT projects at a far lower cost than with traditional cloud methods, and will maintain the scalability and democratization of data.

Industrial Edge and MDP cloud computing can and will work well together! Our Industrial Edge computing is purpose-built and our MDP cloud computing is a more general platform leveraging also DaaS and AaaS, that will work with purpose-built edge systems. 

Feel free to comment here and/or contact me to learn about the possibilities. I’m eager to hear what you think!

Related Tags