Evaluate Weigh the pros and cons of technologies, products and projects you are considering.

Will IIoT edge computing be a real trend in 2019?

Is edge computing analytics a real IoT trend, or is it more smoke from analysts and large technology vendors? As manufacturers like Cisco, Hewlett Packard Enterprise and Dell build specific infrastructure for the edge designed to be more physically rugged and secure, we should all believe there will be a lot of IoT money at the edge.

While working the last few months with Aingura IIoT, I have become aware of the difficulties of developing and implementing edge computing machine leaning in the manufacturing industry. The company combines years of industry experience and knowledge in automation (PLCs, SCADAs, HMIs) and electrical and mechanical engineering with a unique edge computing distributed system used by data scientists to develop machine learning algorithms and build IIoT applications for manufacturing and automotive companies.

However, it is worth asking if clients are ready or even interested in implementing these technologies — or will they continue one more year with pilots that do not go anywhere?

Below are some observations I believe will help accelerate the implementation of edge computing and machine learning in manufacturing.

Get help finding the needle in the haystack

With so many companies talking about Industry 4.0 and with a fragmented ecosystem of IIoT vendors and the challenges that always appear during discussions with customers, it is normal that manufacturers are asked for free pilots.

Source: Aingura IIoT

But it is not just finding the needle (the best or cheapest IIoT offering) in the haystack (the IIoT ecosystem), it is how well this needle matches your business and technology strategies.

I know I am selling myself, but my recommendation is to get advice from independent IIoT experts.

Avoid OT vendor lock-in. We need machine data availability.

Powerful edge analytics machine learning applications need to exchange data with manufacturers’ PLCs. Reading the specifications, one might think this will be easy. In fact, we can find many ways to extract data from PLCs if manufacturers provide info on how to do it. However, most top PLC manufacturers do not allow customers to easily extract data from third parties or their own customers.

It is not a question of protocols; it is a question of vendor lock-in and data availability. Customers must request openness and avoid lock-in if they want innovation in their plants.

Edge computing and machine learning: The last frontier to break between IT/OT

I was optimistic about the quick convergence of IT and operations technology (OT) before. I was wrong. If you visit manufacturing companies’ plant floors, you will see how much work still needs to be done.

Edge analytics is a key component in the integration of IT and OT and requires the combined knowledge of these teams. But the lack of skills in both areas, and the impact in operations and business, makes it difficult to know which department should lead edge analytics projects.

Manufacturing companies need a role with authority, such as a chief IIoT officer, and resources to lead the IT/OT convergence strategy.

To cloud or not to cloud: Don’t let this stop you

When I wrote about fog, cloud and IoT a few years back, the hype around edge computing and machine learning had just started. There was a lot of confusion around fog computing and edge computing and how they would impact the IoT architecture, especially when it came to cloud workloads.

Today, top cloud vendors are offering IoT platforms and tools that combine cloud and edge application development, machine learning and analytics at the edge, governance and end-to-end security. On the OT side, companies like Siemens have launched MindSphere, an open, cloud-based IoT operating system based on the SAP HANA cloud platform.

Manufacturing companies should not stop developing or deploying edge computing and machine learning applications that monitor the health of their machines or improve asset maintenance and quality control because they are afraid of integration with public or hybrid clouds.

IIoT edge computing helps manufacturers improve their competitiveness without the cloud. And when ready for the cloud, it will provide additional benefits, so make sure your IIoT edge system is ready for easy integration.

Connected machines are the only way for new business models

Security is one of the main challenges of IIoT adoption in the manufacturing industry. Manufacturers have been reluctant to open their manufacturing facilities to the internet because of the danger of cyberattacks.

But as we are heading for an economy of platforms and services that need products and machines connected, every factory should be able to tap into machine data remotely and make it available for machine vendors. This requires every edge computing machine learning system implemented to be built with the capability to share data remotely via open and secure protocols and standards, such as MTConnect and OPC-UA.

Having machines connected is the first step to making machines smarter, building smarter factories and flourishing with new business models.

The benefits of using edge computing machine learning systems are very attractive to manufacturers because it allows them to minimize latency, conserve network bandwidth, operate reliably with quick decisions, collect and secure a wide range of data, and move data to the best place for processing with better analysis and insights of local data. The ROI in such IIoT systems is very attractive.

But manufacturers will never achieve these benefits if they do not step up and change their outdated attitude and quickly start their IIoT journeys.

Thanks in advance for your likes and shares!

All IoT Agenda network contributors are responsible for the content and accuracy of their posts. Opinions are of the writers and do not necessarily convey the thoughts of IoT Agenda.

Data Center
Data Management