Industrial automation systems take advantage of remote monitoring and diagnostics for process optimisation and operational efficiency. Increasingly industrial equipment such as robots and plant controllers are being shipped with embedded cellular connectivity. Analyst firm Berg Insight forecasts that M2M connections in industrial automation applications will grow at a CAGR of 23.2% to 7.1m connections by 2018.
The opportunity to generate operational efficiencies extends from machinery on the factory floor, to operations such as mining equipment in remote locations served only by satellite connectivity. It also encompasses partner and supply chain management and has the potential to enable organisations to rapidly arrive at the root cause of inefficiencies or disconnects in processes.
It’s now clear, that if done well, M2M applied to industrial automation can generate improved customer experiences, streamlined processes and cold, hard cash. However, maximising those benefits is dependent on the M2M market continuing to mature, increased standardisation and ongoing efforts to integrate and automate.
M2M Now: What are the benefits to enterprises in terms of increased productivity, reduced downtime and failure avoidance that M2M technologies and services can bring in industrial automation?
David Hofert: There’s now quite a lot of activity underway in the industrial automation market. M2M techniques such as cloud, big data and acquisition of data from sensors are highly relevant to organisations and everyone is moving towards improved processes, efficiency gains and more effective troubleshooting.
Good troubleshooting is more important than ever as Industrial environments are much more complex now. For instance, we’re making multiple, complex products out of different materials including more and more electronics and processes are evolving to reflect that heightened complexity. Sometimes a new process works well and sometimes it doesn’t. M2M capabilities provide a way for organisations to understand why.
It’s important to avoid the enormous mistake in any M2M space of deciding to collect every data point – just because you can – and rely on working out useful insights later. You can certainly collect phenomenal amounts of data but the problem then is to analyse it and get a meaningful result. Organisations must start with the end in mind.
For example, if I have one machine in a particular plant that keeps going down, I need to understand why it fails. Collecting some basic data shows that it always gets too hot just before failure so that’s a clue of where to look further to determine the cause of the problem. It is possible to draw correlations from environmental and machine performance data in the plant and that’s much more of a focused activity than simply connecting – and collecting – everything. There is a virtuous cycle of collecting enough data to understand a process and then reporting information on how the process is running. Make changes, review more data, and repeat.
M2M Now: Which adjacent technologies can be deployed alongside to maximise the benefits? It’s clear that big data analytics will be relevant but how will the value of the data extracted be realised?
DH: There are many disparate influences affecting how M2M solutions and industrial automation are approached. One important technique using the Internet of Things and big data is to try to bring enterprise data processes and IT data management techniques to the physical world. The multiple parties and players involved have to come together in a timely way to achieve that transition.
Oracle, for example, is able to use IT techniques developed over the last 20 years to help you look at information from multiple different areas and to bring it together into an integrated view so that your IT infrastructure can be managed effectively. That’s the same challenge and goal for industrial automation these days.
Of course, industrial automation is a fragmented landscape and each piece of equipment has its own issues. But when you try to understand a plant-wide problem, you can’t determine the root cause from one machine’s input. You need to create a broader way to pull information from a variety of sources and achieve a single view. This is more of a DCA approach than the traditional SCADA-only model. Either way, an important part of the total solution is to collect data and integrate it with your back office systems. Ultimately the data and results need to have a business impact or else it’s all just statistics.
IT has been working on this for a number of years and the opportunity here is for industrial automation to take advantage of the work that has been done in IT for decades.
M2M Now: So the architecture is at least IT-like?
DH: Yes, it’s all about processing and integrating data from disparate sources to provide a holistic view. In industrial situations there’s a concept of islands of automation in which elements of a process are considered separately. For example, car makers will have a series of production lines that ultimately come together and out pops a car. However, shopfloor solutions can’t comprehend the entire plant so you need to be able to understand the data from multiple systems to understand how an issue or change in one segment of the production process causes downstream impacts and affects the quality of the product.
At Oracle, we talk about expressing business intelligence out to the edge. We look at edge nodes – things like management consoles or gateways – to contain a greater amount of intelligence enabling them to make local decisions and report out not just data but information.
Obviously, I can collect the temperature of a part of a machine five times a minute and that’s great because I may need it for deep analysis on occasions. But what I really need to know in a timely manner, is that the machine is operating within set parameters. It’s useful to know if the machine is running properly – this is information built out of multiple data inputs.
To do this the local system needs enough logic to report the edge of a performance boundary. For instance, if the temperature starts to rise it might decide to start sending raw data with the information so the situation can be analysed in depth at the back end as it evolves.
This model offers you local decision-making and locallyproduced information instead of reams and reams of data that is difficult to evaluate.
M2M Now: Industrial automation environments are highly fragmented though. How do you get data from one plant equipment manufacturer to integrate with the systems of a heating and ventilation provider, another equipment maker and potentially several IT systems?
DH: For us, Java is an important aspect of enabling crossplatform communication. It’s clear that more software is needed in everything we touch to enable the benefits we’re discussing. Our value proposition is that, if you’re using Java on platforms and consoles or smart sensors, you have the same programming language as you use in the back office, for big data applications and all of Oracle’s middleware.
There have traditionally been high walls between IT and the plant floor guy and the only way to communicate is to have a bucketful of data thrown over the wall from either side. Java can help break down the barriers and help you decide where that data should go.
With Oracle’s approach we create a platform that stretches from the shop floor to the back office and ultimately to the boardroom. You can migrate logic or applications wherever you need to and wherever makes the most sense because it’s all written in Java and it can run anywhere.
We see Java as a very important aspect of a solution because it means organisations can integrate various teams via a consistent language. They can develop the analytical behaviour of a complete system and evolve it – important since we all know it’s going to change starting day one.
M2M Now: What about the greater bandwidth that 4G/LTE brings? In what ways does the far more granular set of data collected and transmitted deliver real value rather than only the perception of greater value because more aspects are enabled to be monitored?
DH: Wireless broadband communications modules are very important to enabling the concept of sending information rather than data back. When you get past 3G and into 4G/LTE you have very good bandwidth offering you lots of flexibility in terms of moving data and code. Wireless connections also enable solutions where it may be too complex to wire everything together. For example, it could be too costly for a large plant to upgrade to wired industrial Ethernet.
One of our partners, Gemalto, provides wireless modules that run Java ME 8. Those modules provide the flexibility and capability to set up new solutions with existing equipment but use a communications module to run your business logic. It comes back to the model of having a platform where you can flexibly decide what to do and where to do it.
M2M Now: Who are the natural providers of M2M systems for industrial automation? Where does Oracle fit into the value chain?
DH: If you’re starting from scratch to enable your factory for M2M, any number of systems integrators or specialists may be involved. Oracle’s goal is to make platforms – platforms that are robust, secure and scalable because our job is to make it easier for customers and partners to create end solutions.
Providing a rich platform with Java, middleware, database, big data analytics and high transactional capabilities for the industrial automation market provides real value. Why? Because having all those pieces enables people to focus on their business logic and analysis rather than how to connect everything.
The value of having a consistent platform in a consistent language for the guy in the plant on the floor and the guy in headquarters IT is that it provides something solid to build on. M2M and the Internet of Things is ultimately a world of discovery. There is still much more that needs to be understood and optimised.
However, there is value to be achieved today if you think about how to integrate across partners, customers, sites and personnel. Remember that these are very much the same problems that affect any IT architecture and the way to solve them is by achieving a consistent, stable, secure, performant platform.
