How will IIoT discussions evolve in 2018?

Follow @TheManufacturer

As the Industrial Internet of Things (IIoT) continues to evolve, system integrators are beginning to develop smarter, more holistic ways of working.

They have realised that the traditional practice of bidding, designing, implementing, testing and delivering is quite an intensive way of creating revenue in terms of time and labour.

As such, an opportunity exists to develop associated ongoing service offerings, a gap which system integrators are likely to be the best placed to fill. They’ve just implemented the system, they understand how it works and what is required regarding connectivity, cybersecurity and so forth.

They are in the best position to help a customer continue to run that network until they potentially develop their own skillset internally or they decide it’s easier to adopt a service-based proposition.

The Manufacturer recently sat down with John Fryer, senior director of industry solutions at global fault tolerant computer servers and software company Stratus Technologies, to discuss how manufacturers’ approach to IIoT is changing.

Software, platform and infrastructure-as-a-service have all proven popular, but is automation and the industrial internet likely to go down that route? Not necessarily, according to Fryer, for one principle reason.

“There is a subtle difference between information technology and operational technology. From an IT perspective, if email goes down for a couple of hours, it’s inconvenient and people get frustrated, but a business doesn’t grind to a halt, other tasks can still be completed.

“If an OT network unexpectedly goes down, it can have a substantial bottom-line impact to your business, potentially almost instantly. Manufacturers would need to be assured that the as-a-service aspect could address and rectify any issue swiftly, requiring very tight service level agreements. That’s the big unknown, currently; but like everything else, technology and capability will evolve.”

So, how does the IT expert consider digital technology, particularly the industrial internet, to move forward?

“At the beginning of 2016, talk was dominated by IIoT, everything was going to be sent up to the cloud, machine learning was going to take over the world, and everything was going to be connected. That hasn’t happened quite as quickly as many experts and analysts predicted.

“Analytics vendors foretold that IIoT would become mainstream by 2019, a vastly different perspective from those actually operating automation networks. Their conversations are much more focused around cybersecurity.

“As such, there’s been a move towards exploring what can be done from an analytics perspective inside the plant, therefore businesses don’t have to cross that cloud or remote data centre boundary yet. Cost also comes into play here. Cloud computing is often ‘free’ for individuals, but when you start deploying it at an industrial scale, the costs can increase quite rapidly.

“This has led to an increase in momentum and growth around ‘edge’ devices, putting compute power and real-time data analytics right on the production floor and enabling smarter decision-making.

“Though there’s seemingly an ever-increasing number of new terms and technologies, the core concept is still the same – consolidate and virtualise your old networks and if you do that right, it sets you up for when you’re ready to embrace newer, more complex applications.

The role of APM

Asset performance management (APM) has become a major priority for most industrial executives, and it represents a logical first step towards embracing and leveraging the industrial internet.

Fryer explained: “Most production plants collect data, but aside from what is required in the control room, the majority isn’t being interrogated and used. That’s the perfect footing to embrace some of these new ways of working because that’s exactly what you need, the data and analytics.

“The logical next step is to ask, ‘Can APM be performed?’ I associate IIoT in its fullness with machine learning and the use of that kind of technology to do, for example, predictive analytics. Okay, you can do that with APM, but not to the same level of granularity.

“That’s where we are heading over the next 24 months, the first real wave of deployments that haven’t happened to date.”