The need to connect legacy devices to the IIoT
Pre-internet assets lack the connectivity of newer pieces of equipment. These legacy devices, however, still have years of remaining value if they can be linked to the cloud, enabling their data to be analyzed and revealing actionable insights that could perhaps potentially transform business. Wael Elrifai, Sr. Director of Enterprise Solutions at Pentaho, offers insights on how older systems can be made to work with current ones, and talks about the human side of machine learning.
Businesses that have operated for a considerable amount of time will have accumulated several legacy systems over that period. While they have long life-spans, few of these machines will be immediately compatible with one another. The cost of replacing these pre-internet assets to facilitate communication could easily outweigh foreseeable production benefits. What steps must plant managers then take to combine AI capabilities with legacy infrastructure?
— IndIntNow (@IndIntNow) June 1, 2017
Wael Elrifai, Sr. Director of Enterprise Solutions at Pentaho, begins with this premise. “I usually like to remind people that we talk as though data was not generated in the past on these systems. Remember, however, that there’s a lot of robotics involved already, and these systems have sensors that have been producing data for decades. The truth of the matter is that systems such as PLCs (programmable logic controllers) and SCADA (Supervisory Control and Data Acquisition) have already been capturing that data. What you need to do now is to pull the data off those systems. Things like data integration tools are built for that,” he says.
According to Elrifai, predictive maintenance – a technology that drives value in modern manufacturing – isn’t new either. “The difference today is, because the cost of computing and the cost of storage have dramatically reduced, you can do more with it. It’s been a nice positive feedback cycle: Where you can capture more data, you can do more computing work – applied mathematics, machine learning and AI, among others. This then makes capturing data more valuable.”
Conversely, in situations where it’s entirely mechanical and no data or robotics are involved, he suggests looking for proxies for that data. Elrifai adds that while some retrofitting may be required, from his experience a lot of data is already there and is not being used, so it would be best to begin with that.
IIoT implementation across industries
Some industries are more advanced when it comes to IIoT implementation, while others lag. To convince traditional manufacturing companies of the economic benefits of AI investments, Elrifai offers the following examples.
“The ports industry is already using complex machine learning techniques. The most common one for logistics companies is simulated annealing, a method for schedule optimization that sees to it that cranes are doing the right thing at the right time, and containers are moved according to the right schedule.” Elrifai believes that for some container terminals, it’s mostly about the integration of the larger supply chain.
On the other hand, he recalls a visit to a steel factory that wanted to improve its efficiency. “A couple of times a day, they experienced a very specific kind of failure. This cost about 10 percent of their productivity, and in the steel industry that figure is enormous,” Elrifai explains. Furthermore, the way they knew there was a problem was rather unusual: The control room would start shaking.
While the company wanted to reduce this through predictive maintenance techniques, what they didn’t know was that they were capturing all this data already. Elrifai says like many other companies, this particular steel factory would only look at five, ten or 20 variables, the ones that were in their SCADA system.
“What they didn’t do is integrate this with thousands of other sources. The statistical techniques that factories are doing today are low-dimensionality ones,” he continues, adding that trying to convince groups to do more is a matter of explaining to them that it’s just an evolution of what they are already doing.
Where humans fit in
Another dimension to machine learning and AI is the human factor. As far as the supply chain is concerned, Elrifai is of this opinion. “If you are talking about supervised learning – just prediction, basically – oftentimes the baseline data that you use to train these systems is from humans. And you want these systems to evolve, because systems evolve, factories evolve,” he says. “I think humans are always going to be there, helping to state what the ground truth is. Or, at least for the foreseeable future, they will be doing that.”
In addition, Elrifai points out that in certain cases in factories, several different algorithms are voting whether something is going to fail or not, and a human expert is doing that as well.
“With these kinds of methods – ensembles, if you will – you end up with better outcomes. For instance, the machine by itself might produce 75% accuracy and the human on his or her own might produce 68% accuracy. When you put them together, you end up with greater performance, say 80% or 85%,” he states. “I think there’s still a lot of room for cooperation. I don’t think the algorithms are taking over just yet.”
How to solve new problems
“The common problem people have with technology is that they search for problems. That makes no sense. Solution? Start with use cases.”
Finally, Elrifai – with his background in data science – offers this essential piece of advice to companies that plan to connect legacy equipment to the IIoT. “The common problem people have with technology is that they search for problems. That makes no sense,” he emphasizes. His solution? Start with use cases.
“I think there’s a sense that this is extremely expensive to do. However, all you really are doing is putting up a basic data engineering or basic machine learning infrastructure – this is low-cost. There’s a lot of automation available now around machine learning,” states Elrifai. “In the data world, when you try to build models, about 80-90% of the effort that is put in is made up of data engineering, feature engineering, preparing data, filtering – all the easy stuff.”
Elrifai believes that a lot of the data prep for data engineering can be done in an automated fashion. “I don’t think people recognize that. They are trying to use old tools to solve new problems,” he concludes.
Wael Elrifai is an author and speaker. He works as Sr. Director of Enterprise Solutions at Pentaho, a data integration and business analytics company with an enterprise-class, open source-based platform for diverse big data deployments.
Implementing AI in Europe’s Businesses, Beyond the Hype
AI Business set out to find out how AI is transforming business today and how it will evolve in the future. They surveyed the C-Suite executives in the UK & Europe’s 300 largest businesses on how they see AI impacting their organizations, understanding their current and future AI projects, concerns and overall strategy. Georgios Kipouros, Research Director at AI Business, writes about the findings of the survey on techUK.
The majority of the leaders thought AI will transform their industry and saw it essential for their organization. Over 80% compared the impact of AI to that of the Internet. The leaders perceived AI as a way to improve efficiency, reduce overall costs, and also a way to enhance accuracy in their operations. Over 80% of Europe’s leading organizations were investing in machine/deep Learning technologies, expecting to spend an average of 4 million Euros per AI project within the next 2 years.
Read more about implementing AI in Europe’s Businesses at http://www.techuk.org/insights/opinions/item/10724-implementing-ai-in-europe-s-businesses-beyond-the-hype
Keys to effective IIoT design
How does the IIoT change the way industrial products and services are designed? What types of opportunities should companies seize and which challenges should they expect to tackle? Petri Asikainen, Director of Core Technology Development at Konecranes, shares his insights.
Machines report on their usage and condition. Maintenance interaction data is combined with equipment engineering information. Remote-controlled cranes. Logistics chains are truly transparent, as each link is programmed to anticipate the next steps, optimizing flow across the entire chain.
As low-cost sensors and powerful software turn the IIoT from buzzword to reality, the opportunities for new product and service design seem endless. But where should companies start?
Focus on customer value
“The starting point for successful design is a deep understanding of the client’s processes and needs.”
According to Petri Asikainen, Director of Product Development at Konecranes, the key to great IIoT design involves taking a step back from all the cool things that technology can do, and focusing on where it can offer clients the most value.
“Buyers in industrial companies are usually highly rational. Most of them do not care for fancy new gadgets or features just for the novelty of it,” says Asikainen. “The starting point for successful design is a deep understanding of the client’s processes and needs. What is essential and useful for them? Can you make significant improvements in efficiency or work safety? How can you make their life easier?”
Sometimes the solution might not be selling new equipment, but rather retrofitting old equipment with sensors that connect it to intelligent networks to extend its lifespan. Or, it may not mean new sensors but developing software to utilize the sensor data already available in a new way.
“When you have a solid understanding of the user environment, there’s quite a lot you can do by developing the software instead of adding new hardware or sensors – and it’s often more cost-effective too,” he continues.
Design products that communicate
It’s important to note that in the IIoT environment, your products are just small pieces of a larger network. To offer real value, they must fit into, communicate with and improve the client’s existing system.
“With any type of new product, you need to ask the basic questions: How will it communicate with its surroundings and other intelligent systems like enterprise resource planning software? What type of data should it gather and for what end?” explains Asikainen.
Not all data is equally useful either. With increasing amounts of automation, and more smart machines and programs being used at industrial sites, the human operators’ ability to handle new information is often already at its limit. As Asikainen points out, “Especially in user interface design, you should be careful to present only the most crucial information at each given moment.”
Another aspect concerns the high demand for data security in the industrial setting.
“Unlike on the consumer side, you cannot monetize customer data by sharing it with third parties,” Asikainen says. “Data security is a huge issue in IIoT design, as the possible risks with data breaches are often severe.”
Use data to deepen understanding of client needs
“With a clear focus and insight into client operations, IIoT solutions can produce tangible gains for most industrial clients.”
While there are challenges in designing for the IIoT environment, there are also huge opportunities. For one, the influx of new information allows designers to form a deeper understanding of clients’ real needs and to offer them optimized solutions.
“Before, we had an incomplete view into how clients actually use our equipment and services. Now, we can start the design process from facts and real-life data. That is a great advantage for both sales forces and product development.”
In practice, this can mean creating new service concepts or equipment models that fit the clients’ use patterns more precisely. Or it could include offering complementary services that improve safety and quality such as staff training sessions on how to use the equipment efficiently.
The good news, says Asikainen, is that with a clear focus and insight into client operations, IIoT solutions can produce tangible gains for most industrial clients.
“In heavy industrial settings where production volumes and material flows are large, optimizing equipment and tweaking processes with the help of new data can bring in significant savings,” concludes Asikainen.
Petri Asikainen works as Director of Core Technology Development at Konecranes.
Why Industry 4.0 is not only about IoT devices
Commentary on Industry 4.0 that only focuses on the Internet of Things (IoT) entirely misses the point, argues Ben Merton in his article in IoT Tech News. More than 80% of Computer Numerically Controlled machinery is owned by small and medium-sized businesses, who are unlikely to have enough resources to make full use of the IoT functionality of their machines.
The opportunity exists for centralized, technology-driven solutions that use IoT to connect and manage all the entities in a given supply chain. An end-to-end, IoT enabled manufacturing solution will not only reduce the cost and complexity of managing an outsourced supply chain, but also lead to a whole host of environmental and social benefits like reducing the amount of waste, lowering overall costs and localizing.
Read more about the true power and impact of IoT and Industry 4.0: https://www.iottechnews.com/news/2017/apr/18/why-industry-40-not-only-about-iot-devices/
Image credit: oYOo / Shutterstock.com
The Operations Technology (OT) vs. Information Technology (IT) debate turns to better security
While OT managers may see the benefits of IoT-enabled asset monitoring, IT leadership can see IoT connectivity as a security threat. IoT-connected machinery offers uptime rewards at minimal risk but when done wrong, that connectivity into OT systems can pose big threats.
Material Handling Product News has interviewed several security experts on the ways to avoid security vulnerabilities when moving from closed-off OT systems to wireless networks and IoT connectivity.
“Integrating these systems can provide a lot of efficiency and help with goals like uptime, but at the same time, as things become more connected, they become more vulnerable.” says Keith Blodorn, director of program management at ProSoft Technology, which specializes in industrial communications and remote access solutions.
Read more about how companies are solving IoT connectivity data security issues: http://www.mhpn.com/article/the_operations_technology_ot_vs._information_technology_it_debate_turns_to
Image credit: Sergey Nivens / Shutterstock.com
How can Industry 4.0 help the global steel industry achieve greater efficiencies?
Taking place in Warsaw, Poland, the Future Steel Forum assembles speakers from academia and the steel industry to examine how technological innovations can revolutionize steel production. Matthew Moggridge, Editor of Steel Times International, talks about the themes and perspectives steelmakers must consider as they shift to a digital manufacturing platform.
According to the 2016 Global Industry 4.0 Survey conducted by the consulting firm PwC, the buzz surrounding Industry 4.0 has moved on from what some had earlier considered as hype to actual investment and real results. This investment, in turn, is translating into increasingly advanced levels of digitization and integration. 67% of respondents from the metals sector, among them companies in the steel industry, say they expect to reach advanced levels of digitization in their vertical value chains by 2020.
Matthew Moggridge, Editor of Steel Times International, shares a similar view. “The steel industry is well prepared for Industry 4.0 and has, for a long time, been at the forefront of industrial technological development,” he says.
“There are companies, such as Primetals Technologies, SMS group, Danieli Automation, Fives, among others, who have been pushing the boundaries of digital manufacturing and partnering with leading steelmakers such as ArcelorMittal, Tata Steel, Voestalpine and many others to develop the concept of Industry 4.0.”
Moggridge adds that in the US, Big River Steel is arguably the first smart steel plant. The company recently partnered with Noodle.ai, a San Francisco-based Enterprise Artificial Intelligence company, to implement Enterprise AI to optimize operations at the former’s scrap metal recycling and steel production facility in Osceola, Arkansas.
“On the one hand, digitization has moved from being an augmenting capability for steel companies to something that is now becoming a disruptive force. On the other hand, it is delivering supply chain agility, deeper process understanding and higher production utilization.”
Efficiencies and challenges
Broadly speaking, Industry 4.0 assists the global steel industry in its quest for greater efficiencies while raising new concerns. On the one hand, as digitization has moved from being an augmenting capability for steel companies to something that is now becoming a disruptive force, the PwC report says that it is delivering supply chain agility, deeper process understanding and higher production utilization.
The report states: “Automation is combining with data analytics to enable much higher flexibility as well as more efficiency in production. Algorithms are linking the physical properties of the materials with production costs and plant constraints to improve efficiency. Processes that were previously separated are now being integrated, leading to reductions of heat loss, energy consumption, throughput time, inventory as well as better price optimization.”
On the other hand, the people aspect also needs to be addressed. PwC states that companies will need to make sure staff members understand how the company is evolving and how they can be a part of the change. From PwC’s interviews with metals companies, the biggest challenges involve issues such as culture, leadership and the economic case for change.
In addition, Moggridge cites Dirk Schaefer, assistant professor of design engineering at the University of Bath, UK, who argues that the development of a new work force will also prove challenging within the context of Industry 4.0. Schaefer believes that investing in workforce education is essential. “Each of the previous industrial revolutions resulted in a surge of unemployment. There is no reason to believe that this will be any different this time around, unless preventive action is taken today,” Schaefer asserts.
— IndIntNow (@IndIntNow) June 14, 2017
These topics will be addressed by experts at the Future Steel Forum in Warsaw, which takes place today and tomorrow. Other discussion points include the impact of smart manufacturing on the steel industry, Industry 4.0 and its implications for plant safety, the future of cooperation between automation and steel manufacturing, and the role of human beings in the factory of the future.
Taking part are speakers from academia, the steel industry and the world of steel production technology such as Dr. Rizwan A Janjua, Head of Technology, World Steel Association; Jose Favilla, Director, Industry Solutions for Industrial Products, IBM; and Professor Chris Hankin, Imperial College London, among others.
Connecting the dots
As for Industry 4.0-related themes that are set to gain ever greater prominence in the coming years, Moggridge, who will deliver the welcoming and closing remarks at the event, has this to share. “Cyber security will always be a big issue that will constantly need to be addressed, but also the role of the human being in an increasingly automated environment, not only in steel but in other areas of industry as well,” he says.
“What people tend to forget about the steel industry is that it is already a very automated environment. In many ways, it’s just a case of connecting the dots before steelmakers can claim to be true advocates of Industry 4.0.”
Matthew Moggridge is the Editor of Steel Times International. The Future Steel Forum takes place in Warsaw, Poland, on June 14–15, 2017. futuresteelforum.com
How the Internet of Things impacts supply chains
Enterprise resource planning and supply chain management (SCM) have gone hand-in-hand for quite some time, but the IoT revolution will allow those solutions to be enhanced by intelligently connecting people, processes, data, and things via devices and sensors.
“Think of it as SCM 2.0,” writes Udaya Shankar, Vice President and Head of Internet of Things for Xchanging, a business process service provider for the global insurance industry. According to Shankar’s article in Inbound Logistics, this deeper intelligence can come to life in many different ways when it comes to supply chain data and intelligence – from the automation of the manufacturing process to improved visibility within the warehouse.
One area that Shankar believes will play a prominent role in the future supply chain, as it’s impacted by IoT, is in-transit visibility. “The logistics ecosystem has many players, and thus, many moving parts. Products are handled and transferred between the manufacturer, suppliers, the distribution center, retailer, and customer.”
Read more about how IoT can help supply chain professionals at:
Image credit: Lightspring / Shutterstock.com
The deciding factor – how to utilize IoT data analytics for business intelligence
To make the most of data, it has to be transformed into information, which then has to be transformed into intelligence. As companies seek to leverage data – whether it’s internal, external, structured, or unstructured – to improve profitability or boost operational efficiency, analytics makes it possible to gain insights on business areas that were previously out of reach. Alun Jones, Data Scientist at Konecranes, talks about how organizations can best use IoT data analytics to arrive at more impactful business decisions.
According to McKinsey, the potential economic impact of the IoT could reach $11 trillion per year in 2025. That figure is equivalent to around 11 percent of the global economy. Turning that possibility into reality depends on how effectively IoT data analytics is used to drive better decision-making. The technology research firm Gartner identified IoT data analytics as one of the key IoT-related technologies that should be on every organization’s radar in 2017 and 2018, second only to security.
To identify target-rich, high-value data that can be used to generate business intelligence, the following steps should be taken.
1. Be aware of what you already have. It makes sense to know if some of that information is already available or accessible even if it isn’t immediately apparent. If you don’t know, then find out: Build a data map for your enterprise.
2. Think like a custodian, not an owner. The term “data owner” can be misleading as it appears to imply not only ultimate responsibility, but also the ability to utilize data for one’s own purposes. Both are not necessarily true of data use within a business. A data custodian, meanwhile, is responsible for the technical environment and controls around data.
3. Every action is part of the value chain. The siloed approach to data access makes unifying the analytics layer a challenge. To generate scalability and real-time performance, however, all types of analytics – descriptive, diagnostics, predictive and prescriptive – must be brought together into a single engine.
The role of cloud analytics platforms
In terms of using cloud analytics platforms to derive value from IoT data, it’s important to remember that not all data is created equal. Companies should think of ways to get data from a device into a position where it can be analyzed; the priorities of that data need to be determined as well.
Next, it’s also essential to gather data from numerous sources as interoperability is key in a heterogeneous environment. Last, it is advisable to have distributed data sources so that the cloud is there by default. Cloud simply means off premises; there will be distance between individual data sources and the computer power performing the processing. If you are uncomfortable with the cloud then find out why, and work to alleviate those anxieties. Processing IoT data close to the source results in less network delay than transferring it to the cloud, processing it there, and sending back the actionable result, such as computing and analysis at the edge.
“Harnessing IoT data analytics for business intelligence is not a one-time exercise, but a continuous process.”
As far as the barriers to widespread IoT value delivery are concerned, these could be overcome in two ways. First is technical. This covers everything from data gathering and low power or no power devices. (At the moment, for example, sensors and devices need power to drive them or the transmission of data, and in the future there will be a need to have devices that have lower power requirements.) Data architecture and cost of hardware should likewise be considered.
Second is the people aspect. Gatekeepers need to change. Management must improve its ability to understand and interpret the output from analytics. Individuals need to collaborate, even with those outside their respective enterprises. Normal business practices mean that things are driven on short-term departmental measures – this must be reconsidered as well. Do you design your plant to be cheaper to build, or more efficient and flexible to run?
Overall, harnessing IoT data analytics for business intelligence is not a one-time exercise. Rather, it’s a continuous process. Bear in mind that not everything is going to work. Optimizing what you do today is not enough either.
In addition, look at how to change the business model in a way that fits the market. For instance, GE builds airplane engines. They innovate by fitting sensors to gather that data and transfer it back to the factory. This is adopted over time so all engines send data about themselves. Over time, this is optimized so that GE knows the state of each asset and is able to predict when parts are likely to fail. This reduces downtime, making maintenance more efficient. Once the asset behavior is understood and de-risked, the business is transformed from building engines to offering engines as a service. GE’s software platform is now the key element in their business model. Cranes are a little behind but are catching up fast as the platforms needed to support such devices are already being built.
Alun Jones works as a Data Scientist at Konecranes. He is participating in several panel discussions at the IoT Tech Expo Europe event in Berlin on June 1-2, 2017.
— IndIntNow (@IndIntNow) June 7, 2017
Learning by doing and the value of pivoting in an IIoT environment
Data-driven technologies and business models are becoming increasingly linked. For many companies, this means that the skills and assets they are today investing in bear little relation to those they sought five years ago, and neither do the services they offer. Mika Parikka, Managing Consultant at LINK Design and Development Oy, holds that the Industrial Internet of Things plays a central role in this redefinition process, as it is fundamentally affecting how businesses are being run and what kind of talent is deemed necessary.
According to Mika Parikka, the Industrial Internet has introduced a distinct shift in business models from times past, when often what companies sold was technology or machinery itself. Parikka is the former CEO of TreLab, an electrical manufacturing company whose activities include retrofitting wireless smart sensor solutions for industrial equipment.
“The question of IIoT must be understood as not being primarily about technology, but instead about positioning and differentiation. After you understand how the machinery is being used, you can begin to see how the business is, or should be, run,” Parikka says.
He doesn’t see the Industrial Internet as being completely defined by technology, although it determines the types of data companies will be able to offer their customers. Herein lies the crux to his argument; customers will essentially come to define the businesses which serve them. Parikka maintains that for companies to reap the benefits of IIoT and IoT, mentalities need to shift form ‘we are only selling hardware,’ to ‘we are selling services that benefit our customers.’
“This affects who the products and services are being marketed to, as well as how contracts between parties are being drawn up,” Parikka says. “At TreLab we provided our customers with technology which they could utilize in IoT and IIoT scenarios. We began seeing the kinds of problems that customers were having in selling and conceptualizing their services as they began telling us that their businesses were fundamentally changing, and that they are no longer billing their customers for what they used to.”
This is not a surprise when considering the types of customer relations that the Industrial Internet is spawning. “Previous business models were largely about selling machinery and technology to customers, whereas in the future, they will largely be defined by customers wanting total solutions. Thus, the most significant change that will occur is that sales, customer services and most notably contracts, will need to be done from the perspective of customer value, as opposed to selling things in the traditional sense.”
“The question of IIoT must be understood as not being primarily about technology, but instead about positioning and differentiation. After you understand how the machinery is being used, you can begin to see how the business is, or should be, run.”
Trial and error yields the best results
Parikka cites Kati Hagros, the Chief Digital Officer at Aalto University and former CIO of KONE, as someone who often speaks about the need for companies to learn from their customers.
“She often urges companies to go with the customer and try things out. Then after these tryouts, once the business model begins to take shape, to fund it lavishly so that is has the ability to overcome corporate inertia, which tends to stifle ideas that appear radical at first,” Parikka relates.
The willingness to engage in these tryouts necessitates a significant internal philosophical shift in companies, most of all from the people at top management level. However, Parikka feels that a mentality shift shouldn’t be seen as totally restructuring what a company has been doing or selling up until that point. This is particularly true for larger organizations who have the resources for testing and experimentation, as well as for setting up smaller units that have more autonomy in terms of serving the customer.
“I believe that those at top management level must start acting more like venture capitalists. Potential solutions to customer needs exist, particularly among those at the frontlines of organizations who understand and see both problems and solutions. The role of top management is to then recognize and understand which of the possible solutions to fund to put the company in the best possible position, and to allow the frontline people lead the change,” Parikka says
He asserts that this is particularly true for industrial organizations who manufacture and produce material products. “Especially for organizations that are now moving into providing anything as a service, the leadership need to clearly lead the way, saying: ‘this is important in our business, we need to do this and above all else try that.’ It is important to prioritize what seems to be the best option, going for it, and changing direction – or pivoting as they say in the start-up world – if it doesn’t work. Learning by doing is definitely the way to go.”
The birth of the knowledge worker
After successful learning and experimentation has taken place, the emerging business models will set new demands upon the workforce, breathing new life into them. Drawing on the topic of companies requiring new expertise and new talent, Parikka highlights how knowledge workers will play an essential role from an operational standpoint.
“Most of the work will be about defining the procedures and algorithms, when there are a lot of sensors and subsequent data coming in from different sources. A knowledge worker will need to be able to analyze vast quantities of data, and really understand where all the data comes from. On top of this, the worker will need to create ways to deduce action-oriented alerts from the data. In most cases this is not what we would call programming, but instead, the type of work we would be doing every day in Excel – creating rules and procedures to act on.”
Parikka, who is an engineer by trade, holds that this workforce will be increasingly made up of people with a technological prowess, who will then take the lead when it comes to business initiatives. He concurs with Bill Gates’ statement that ‘it’s very difficult to teach technology to business people, and thus much easier to teach business to technology people.’ Approaching business from a technological mindset will mean that issues such as IIoT data ownership will be prioritized and dealt with early on in contracts signed between service or data providers and their customers. For tech people to spearhead business model realization processes, they ultimately need the license, capacity and funding to experiment to get the best results.
Image credit: Rawpixel.com / Shutterstock.com
The future of IoT and machine learning – what role will humans play?
Despite having been around for over two decades, machine learning and its integration into business models is yet to become commonplace. Jari Salminen, Managing Director of Cumulocity, has witnessed the unfolding of machine learning and noted the progress made in its adoption into a wide range of industries over recent years. He says that rather than spend time in building grand strategies then assume that value can immediately be realized, companies should take pragmatic steps to connect their assets and start collecting data.
“What we are seeing today is that there typically exists a bit of a delay when companies start connecting assets and collecting information to be able to rely on machine learning algorithms and their accuracy,” says Salminen. “The training of these algorithms requires large amounts of data and thus time. It takes time for any individual company to move through the cycle of starting with very basic use cases and moving onto more complex algorithms and dependencies, and eventually introducing machine learning.”
At Cumulocity, Salminen deals with several manufacturing and industrial companies. He recognizes companies that require warehouses – or those whose supply chains do – currently expect sophisticated IoT solutions from a production and manufacturing point of view.
“Things are changing at such a pace that it is now very cost efficient even for smaller companies to deploy off-the-shelf IoT solutions for their supply chains.”
Salminen encourages companies who have examined the cost of IoT solutions for manufacturing or supply chain management over recent years to do so again. “Things are changing at such a pace that it is now very cost efficient even for smaller companies to deploy off-the-shelf IoT solutions for their supply chains as the price of hardware, connectivity and software has dramatically reduced over the last 5 years,” he reasons.
Holding algorithms accountable
The accuracy and efficiency at which these solutions can be implemented will greatly depend on the algorithms triggering them. According to Salminen, the more automated these algorithms become, the greater influence they will have not only on supply chains, but beyond them as well.
“Today, most actions are still done by human users and there are several reasons why that is. For example, accountability – in the long-term, we will need to ask how this will change. Will decisions be made based solely on the algorithms that machine learning will make possible? Maybe. However, such questions and their answers go beyond technology itself, as they are concerned with making sure that somebody other than a computer takes responsibility for decisions. It’s a complex domain,” Salminen says.
So, how to legally and ethically approach decision-making when no human is involved in the process? Very similar questions are currently being asked in the automotive industry.
Closely tied to the accountability of machine-to-machine decision-making within IoT projects is the matter that is security. In general, issues surrounding security are never too far away when connectivity, machines and material networks are concerned.
Multilevel security management
“The issue with security is that it exists on so many levels in IoT projects: from hardware – meaning devices, machines and assets – and connectivity, whether that entails using mobile connections or new narrowband IoT solutions, to the backend, meaning the cloud or servers. On top of all that you might have special applications for company users, partners or even consumers. Security needs to be controlled and monitored at all these levels.”
“Often, security breaches take place when more than one area has been overlooked. A most common area attracting hackers and attacks is wherever any connected device exposes ports reachable from the internet. These are being scanned by hackers,” Salminen continues. He reminds companies that if they are dealing with hardware, that they should keep their wide area network ports closed; from a connectivity perspective, they must make sure that everything is fully encrypted in transit.
“If you are providing cloud services on top of hardware and connectivity, make sure that your whole system is robust and secure. Ultimately, there is no single thing companies must consider when it comes to security, but rather, it is an area that must be owned end-to-end by someone in the project.”
Setting future standards
When asked how he sees the future of machine learning and IoT taking shape, Salminen is simultaneously restrained and excited. He sees industries as being past the initial hype. The implementation of more advanced and sophisticated use cases is now becoming a reality.
“The connectivity costs per device are decreasing, and thus enabling many new developments when it comes to industrial solutions. This is happening across widespread assets, meaning that for example, narrowband technologies are becoming more popular. However, what is still lacking from IoT and will need to be addressed in the next few years is standardization.”
While Salminen is adamant that a comprehensive philosophy regarding standardization must be adopted, he does not foresee the use of one single underlying standard that will apply across industries.
“My guess is that there won’t be an overarching standard that will include everything from devices to data structures. There will be so many different IoT use cases that it will be impossible to create something that would cover all of them,” he says. “We see lot of traction with MQTT (Message Queuing Telemetry Transport) as a messaging protocol due to the fact it doesn’t even try to standardize all parts of the IoT stack. For example, it does not deal with message payload format which is left to the developer to decide.”
On the other hand, Salminen believes that standards like Lightweight M2M by Open Mobile Alliance are not being picked up by market players because probably it doesn’t fit many use cases, among other reasons.
Nonetheless, he reminds those looking to initiate IoT projects to certainly consider available standards, but not to be limited by them. He also states that the key thing to ensure is that you are not locked into any standard, but to have flexibility in case your needs change in the future.
“If I were starting an IoT project, I would be looking at the most recent connectivity options, the role of standards, if such exist, and whether or not they are relevant to my business. However, I wouldn’t force any standards at the moment, as there are in fact very few that are relevant,” Salminen concludes.
Image credit: ImageFlow/ Shutterstock.com