Edge computing is a real buzzword these days. Ask a random IT-company: „Do you do edge computing?“, and the most likely answer will be yes. The term is so overused, in many cases, it has become close to meaningless, stating nothing more than: „Yes, we have an industrial PC somewhere, and there is a program running.“ Well, that doesn’t quite catch the essence of edge computing, and it certainly falls short of covering its full innovative potential, most notably for industrial applications.
The era of cloud computing
Looking back at the history of information technology we can make out several waves of local vs. centralized data processing. In recent years, we have seen a strong trend towards cloud computing. Data management and computing are being moved to large centralized server farms, and the many personal computers are reduced to hosting a web browser.
The advantages of cloud-based data management are apparent:
- easy quick software updates (as they can be handled from a centralized source)
- a global and integrated view of all workflows and f cloud-connected equipment enabling both global and local decisions
- one central place with all data for subsequent optimizations
Siemens has played a major part in the implementation of this cloud-computing trend into industrial settings. MindSphere, the IoT operating system that connects physical cloud environments to the IT-world, was launched together with MindConnect, the gateway that connects industrial equipment, worldwide fleets, and entire plants to the IoT. This has opened the door for cloud-based analytics and better decision making in the industry.
The limitations of cloud-based data management
However, this bright new future hasn’t come without its downside. With the general transition to IoT, cloud-connected machinery and logistics, we also encounter some new challenges. In a time when real things like cars, buildings, and turbines are equipped with virtual doppelgängers, their digital twins, more and more data is being collected to be processed within cloud-based data management applications.
Bad vibes can ruin your work day. We all know that. But bad vibrations aren’t exclusive to humans. When vibrations in a piece of machinery run out of control this can easily lead to a chain reaction disabling a whole production line. So modern control systems come with an ever growing amount of sensors implemented into crucial parts of machinery. They detect these vibrations and send their data to the cloud-based operating system for constant analysis and e.g., predictive maintenance. Imagine all the little sensors producing data within an industrial production setting. The data load is enormous.
This is where we encounter one of the chief challenges for cloud-based data management in industrial applications. If you are involved in cloud-based data processing, chances are these concerns are also yours:
- physical limitations of data transfer (even bits and bytes can’t travel faster than the speed of light!)
- dependence on network availability (when the network breaks down, cloud-based motion control and optimization also breaks down.)
- data load (even at high speed data transfer of this magnitude is too slow)
- data privacy (there is data industrial companies don’t feel comfortable to share with the cloud)
- cyber security (data transfer within cloud-systems always entails vulnerability for data theft; that’s why data transfer within cloud-operated systems is usually confined to upstreaming, as downstream transfer is even more sensitive to cyber attacks)
Handling a heavy data load is easier when data is processed locally. So are we set for a revival? Is the tide of IT-history turning for a new wave of local data processing? Yes and no. Enter edge computing.
Enter industrial edge computing
Edge computing is not new. Major IT-players like Cisco have employed it for years together with its sibling from the fog. The true innovation lies in its most recent integration with industrial production processes and their optimizations within cloud-based systems.
Not only do edge applications like Analyze MyWorkpiece offer the possibility of collecting and analyzing data close to where it originates within the production process, at Siemens they are also integrated with MindSphere, the cloud-based IoT operating system. In edge computing data processing is not exclusive to the cloud’s core, it chiefly happens on the periphery, the edge of the internet where it makes contact with the physical world.
The integration of edge computing within industrial clouds enables us to stick with the benefits of cloud-based systems, like quick and easy software updates while, at the same time, profiting from the advantages of local data processing, like data security and quick reactions of control applications and environment within an industrial production process.
With Siemens Industrial Edge the company has created an open platform for SINUMERIK and SIMATIC Edge applications (SINUMERIK Edge and SIMATIC Edge are digitalization platforms as instances of Siemens Industrial Edge). It integrates with these control systems and extends them with the possibilities of edge computing. Optimize MyMachining / Trochoidal and Analyze MyWorkpiece are just two examples of SINUMERIK Edge applications created within Siemens Industrial Edge. The platform is not exclusive, it is open to all machine users, machine builders, and external app developers. You can create your own edge applications customized to your or your customers’ needs within Siemens Industrial Edge and in the future even use the platform for sales distribution. Think of it as your App Store for industrial edge applications.
So it is not just about letting the tide turn back in favor of the local. It is about reaching a middle ground, a new balance that combines the best of two worlds, local data processing and the centralized cloud-operated systems. Siemens is a leader in this emerging field of industrial edge computing. It’s not just some random program running somewhere on an industrial PC, Siemens offers a fully integrated system for remote management, that provides a working balance of OT (automation, technological integration, data security) and IT requirements.
Chief concerns and how industrial edge computing is the solution
Whenever I talk to clients about these new industrial edge applications, I am astonished by the level of enthusiasm I encounter. They are thrilled. Demand is immediate. Usually, it takes a little more persuasion and concrete cost-efficiency calculations for clients to buy into an idea or a new IT-technology. Common concerns are:
- implementation risks of digitalization
- purchase of a niche solution with low scalability
- purchase of a prototype that doesn’t work for all machinery
Presenting them with industrial edge applications it is easy to dispel these concerns. They offer fully integrated, implementable, and cost-efficient solutions for one of the most imminent issues industrial production managers are facing today: How do I keep my data private, cope with a soaring data load, reach improvements promised by digitalization and profit from cloud-based data management at the same time?
Industrial edge computing brings real innovation to the field. A true union of opposites, it bridges the gap between cloud connectivity and edge management and reconciles the rift between the local and the cloud. The industry has now delved into product development for integrated edge cloud solutions. I am eager to play my part.
A glimpse into the future: industrial edge and AI-integration
You ask yourself how AI fits into this picture? Well, with the integration of MindSphere and industrial edge computing we now have the optimal setup to get AI and machine learning involved. That’s a whole new story and a whole new treasure trove of future innovation opening before our eyes.
Feel free to comment and share your ideas on this matter! Or ask for more detailed information about industrial edge computing and AI-integration.