As the Internet of Things (IoT) becomes more ubiquitous throughout the supply chain, more useful data is being captured. The challenge is applying the intelligence and analytics to produce measurable results.
Some companies have been successfully doing it for decades, but on a smaller scale. UPS and other major distributors have used barcodes, ruggedized handheld scanners and wireless networks to optimize their supply chains and improve service. Beyond distribution channels, however, IoT is seeing new life on the manufacturing floor and even in our homes. “IoT technology is revolutionizing supply chain management across many different vertical markets, allowing improvements in efficiency, productivity and profitability,” says Chris Moorehead, IoT sales for Gemalto NORAM. “Smart manufacturing provides full visibility of assets, processes, resources and products, allowing companies to more closely match supply and demand. This in turn supports streamlined business operations, optimized productivity and improved ROI.”
According to IDC data, published in early 2017, the manufacturing segment invested $178 billion in IoT spending in 2016, twice as much as the transportation segment—the second largest IoT vertical market. Gartner predicts that by 2020, IoT device shipments for manufacturing will reach 358 million units with $160 billion in spending; that’s a 34 percent CAGR in just five years.
Moorehead explains that both the challenge and key to success in smart manufacturing is connecting equipment, integrating diverse industrial data and securing industrial systems for the entire lifespan of equipment.
“Successful companies are using IoT systems and data to make informed changes to their processes that result in better products, reduced development time and resources, more effective customer service and improved security,” he says. “The most successful companies are using embedded systems and cloud-based application enablement solutions to instantly analyze data to drive real-time actions. Results are measured in terms of cost and time savings, improved productivity and efficiency, more effective processes and increased profitability. The reach of the IoT is vast and the possibilities are truly endless.”
Data Types
The type of IoT data being used throughout the supply chain varies by vertical market, but can include energy consumption and production data as well as GPS data temperature, tracking and vibration.
“The types of data collected and utilized are determined by the user based on what is most beneficial to their business optimization,” says Moorehead. “In supply chain management, IoT data are being used to improve not just this field but also the entire company’s ecosystem and its interaction together—from design, to manufacturing, to distribution and logistics, all the way to sales.”
In design, he says, companies like Amazon are using social data and ordering histories to better understand product requirements. They also can use the data collected throughout the development process, like sourcing data, to identify which components and manufacturing variables led to the highest quality and/or lowest priced products.
In the manufacturing plant, data are collected from sensors throughout the floor to analyze operational performance and determine the optimal parameters of speed, temperature, maintenance and the like to maximize quality, yield and keep manufacturing lines running at peak performance.
Distributors have been using IoT data from fleet management and monitoring solutions to dynamically re-route trucks and improve distribution networks, saving time and money, while keeping the end customer in the loop on when they can expect their shipments.
And smart homes, smart cars and digital retailers like Amazon are using data from connected devices like Alexa and Amazon Echo to better understand consumer needs and predict demand.
“There is great potential for optimization as a result of the convergence of IoT data from smart manufacturing, supply chain management, digital retailers and the smart home and smart car scenarios,” Moorhead adds.
Data Collection
IoT data are collected through sensors and systems and sent to cloud-based applications that analyze it and transform it into meaningful information that companies and individuals can use to make better decisions. Data are indexed, analyzed and stored at various points throughout the ecosystem—in embedded system devices collecting data, in the cloud within IoT enablement platforms that receive data and merge it with contextual data for better decision-making, and in backend systems that receive IoT data.
Redwood Logistics, a full-service transportation brokerage that specializes in moving freight throughout North America, has built its business model around data movement—whether it be from an IoT-enhanced device or an onboard electronic logging device (ELD).
“We've taken an IoT data pipeline approach to everything we do, says Redwood Logistics’ CIO Eric Rempel. “We’ve reengineered our entire connectivity platform to an API-led architecture, where we can receive and connect data from anywhere and to anything, and enrich that data as well as pump it into a data warehouse.”
In the past, he adds, most business’ data warehouse endeavors have been more of a data swamp than a data lake because the data just sits there.
“You might be dumping all of your data into the cloud bucket or into a database, but the question is: Do you know how to access it? And is it performant, so that if someone has an analytical business question, you can answer it quickly and easily without the IT department getting involved? Rempel adds, “We're trying to move all of that data away from IT to line-of-business users.”
And while Redwood Logistics does not leverage a lot of IoT devices (IoT devices are probably sending in about less than 5 percent of the data it collects), the end goal is having the ability to capture data no matter the source and index it, while also replaying and enriching it later at high speeds.
There are several challenges to this approach, Moorehead notes. Chief among them is developing systems with the longevity to meet the IoT business case of staying operational for 7 or more years.
“Devices need to be built to a higher quality standard than consumer products, and these end devices need to have field diagnostic and OTA (over-the-air) capabilities to extend their lives,” he says. “IoT solutions also need to be prepared for wireless network evolution, so that when one network sunsets—for example, when 2G networks cease, and 3G and 4G networks take over—the devices need to remain operational without a site visit and hardware replacement.”
Security Risks
Arguably, the biggest challenge facing IoT data today, however, is one of security.
“We need to be able to keep data safe, and ensure its integrity throughout the entire ecosystem over the lifetime of IoT applications,” says Moorehead.
The IoT marketplace has already realized the importance of building trust into the ecosystem, which include all elements of an IoT system: the device, the network, the data and the cloud. Moorehead recommends the following four fundamental elements to effectively secure all IoT applications and vertical markets—from smart cities to supply chains to smart home appliances:
- Authentication/identification—Each device in the IoT ecosystem needs to be able to identify itself and prove its entitlement to access the system.
- Confidentiality—Data transmitted must be encrypted effectively, ensuring it has no value to anyone stealing it.
- Integrity—Companies must ensure that what is sent is what is meant to be sent.
- Non-repudiation—Companies must have incontrovertible proof of the validity and origin of all data transmitted.
Over the past few years, IoT products and processes have progressively gained momentum in nearly all sectors. Therefore, the risks associated with IoT also have increased. At Allianz Global Corporate & Specialty (AGCS), a large part of the insurance provider’s job is to underwrite, examine and prevent said risks, explains Jenny Soubra, U.S. Head of Cyber, Tech E&O at AGCS.
“IoT, specifically, generates an exponential amount of information. One of the biggest challenges is that many companies cannot quantify how much data they have and where inside or outside their network it is stored, accessed or passed through,” says Soubra. “Additionally, a manufacturer, for example, may have multiple data feeds from different sensor-embedded automation technologies that need to be integrated with each other. We analyze how organized and available the data is, if it’s personally identifiable information (like Fitbits and health data), what potential liability is involved if it’s unsecure or under protected, and what damage may be done if the technology malfunctions, such as a manufacturer shutting down its entire operations for a week versus a manufacturing/production day of five hours.”
Soubra says AGCS encourages its clients to utilize integrative technology platforms and consultants, such as dashboard management, to sync all processes most concisely. She adds that companies should be looking to see how these platforms protect data, implement security focus, have an override if there’s an attack, and utilize proper shut down procedures via disaster recovery and incident management plans.
For example, in October 2016, a Dyn cyberattack caused a massive denial of service for domain name systems, rendering platforms, services and networks inaccessible. Hackers attacked by hacking into millions of IoT devices—like smart refrigerators, printers and baby monitors—via a botnet. They took over these IoT devices and used them to send domain name lookup requests to overload certain networks, making them inaccessible.
The IoT manufacturers now may potentially have liability issues for creating unsecure products, like baby monitors with default passwords of “admin.”
“We consequently like to see that these types of companies are leveraging technology to ensure that the devices cannot be attacked remotely again, looking at their security and technology processes,” explains Soubra.
There are a number of technological solutions by way of risk management, which include automated technology that alerts manufacturing companies that the default passwords on a smart refrigerator have not been changed in over two months and may now ping the customer with that information to remind them to change their password, or the use of cloud dashboards to monitor devices and ensure there are kill switch and override technology to avoid manufacturing delays and damage.
Most popular now, Soubra says, is the potential to integrate and leverage machine learning/artificial intelligence into this process to understand patterns. A high influx of activity at one specific time on a smart refrigerator product, for example, may signify a potential device hack attack. Machine learningallows companies to identify this in real time, or even preemptively, and implement appropriate solutions like the kill switch or alerts.
“Companies with a strong understanding of the storage, nature and integration of data are most successful in producing measurable, actionable results from both a product and security perspective,” Soubra adds.
