By Tom Haunert
Smart connected devices are a good start, but businesses can get better results by using device-generated information across multiple applications to drive change. IoT applications for different types of devices and operations—from industrial devices and manufacturing to transportation and field service—can also use device data to generate and share insights in ERP, supply chain, and customer experience applications.
Oracle Magazine sat down with Oracle Group Vice President IoT and Blockchain Applications Development Bhagat Nainani to talk about today’s enterprise IoT, the applications approach to IoT, and the power of IoT to drive connected business outcomes.
Oracle Magazine: What is the state of enterprise IoT?
Nainani: In the last five to seven years, IoT has gone from being sort of a buzzword or a cool new technology to being something that businesses are starting to deploy and get business value out of. Having said that, I think the rate of enterprise IoT deployment or growth has been a little slower than expected or predicted. So, yes, there are real IoT deployments, but a lot of projects get stalled in the proof-of-concept or pilot stage.
IoT projects that are much more likely to be successful start with a business-outcome-based approach.”
IoT projects that are much more likely to be successful start with a business outcome-based approach, where there are clear KPIs [key performance indicators] that businesses are trying to improve with IoT. Next, those projects look at what kind of analytics they need or what kind of data sources they need. And then those IoT projects focus on where that data comes from—systems, devices, and so on.
Oracle Magazine: How are organizations looking at IoT application choices? What are Oracle’s current IoT offerings?
Nainani: There are many vendors that offer a lot of IoT platform components and tool kits. They offer platform components to get your devices connected and to apply business rules. And then they offer AI and machine learning algorithms that businesses can use to create their own IoT solutions.
But businesses—industrial businesses in particular—really don’t want to assemble these things themselves. They would like solutions that they can readily deploy to solve their specific business problems. They would prefer to use an asset tracking solution, for example, which they can just configure with their assets and extend it as needed. We’ve seen this trend at Oracle, and over the last couple of years, we’ve shifted focus to more purpose-built IoT applications for addressing specific domains.
We provide five IoT applications today. Oracle Internet of Things Asset Monitoring tracks assets, their utilization, and their service needs. Oracle Internet of Things Production Monitoring looks at the factory floor, performs diagnostics, and measures performance against plans. Oracle Internet of Things Fleet Monitoring monitors trucks, shipments, and driver behavior. Oracle Internet of Things Connected Worker tracks the health and safety of industrial workers. Oracle Internet of Things Service Monitoring for Connected Assets Cloud Service runs remote diagnostics for field service.
And these applications are built on our Oracle Internet of Things platform components, and they integrate with Oracle’s ERP, supply chain, and customer experience apps. We effectively extend our business applications with insights and data from our IoT applications.
Oracle Magazine: What are best-practice strategies for handling the volumes of IoT data and getting value from all of that data?
Nainani: Less than 1% of IoT data is actually used today. Some companies start by getting basic alerts from IoT devices when something has crossed a certain threshold value. But a lot of devices have those capabilities built in.
But the real value of IoT comes from getting insights from the devices—from analyzing the data streams in real time and using machine learning to get those insights.
We look at a business-outcome-based approach for applying machine learning and AI to IoT data.”
We look at a business-outcome-based approach for applying machine learning and AI to IoT data. First, employ KPIs to measure the outcomes; they can include equipment efficiency, asset utilization, production yield, and so on. Next, identify anomalies and use machine learning algorithms as part of the identification. And again using machine learning, create predictions based on the anomalies reported, such as “this machine will likely fail before its next maintenance.” Next, utilize AI for prescriptive actions and root-cause analysis such as “this part is the likely cause of the potential failure” or “the system recommends that this specific machine be maintained earlier than its planned maintenance window.” Finally, apply forecasting to look ahead to optimize operations, such as using production-line device information to help order the right quantity of raw material.
Because IoT information is used across applications, you need to connect IoT-based insights back to your applications to drive change. For example, when a vehicle failure means a truck can’t deliver on schedule the material needed for a field service repair, both the logistics application and the field service applications need that information. That kind of information flow is called a digital thread, where IoT data gets used in the context of many other applications and provides a bidirectional workflow that connects the applications.
LEARN more about Oracle Internet of Things.
LEARN more about Oracle IoT applications.
Photography by Oracle