TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

Q&A: Untangling the Potential in the Internet of Things

Dell's general manager for advanced analytics explains some of the many uses and tremendous benefits he sees ahead for the IoT.

• By Linda L. Briggs
• January 12, 2016

The Internet of Things (IoT) is generating interest and excitement and for good reason. Hardware, software, and connectivity have come together to make networking devices possible in new ways. What is the IoT and how can businesses use it? Where does the value lie and what are some use cases? In this interview, John Thompson, Dell's general manager for advanced analytics, explains the basics of the IoT and points out some of the possibilities that lie ahead. For example, "there's a tremendous need to make the world greener and I think the IoT is going to be one of the great drivers of that," Thompson says.

BI This Week: Organizations are trying to understand the IoT and how they can use it. Let's start with a basic definition -- what is the Internet of Things and how does it work? How is it different from the older concept of machines connected to the Internet, which we've had for a while?

John Thompson: The term IoT itself is fairly new and is attracting plenty of buzz. As you said, the predecessor concept is machine-to-machine connectivity. That refers to an older, industrial environment with a production-line stamping machine, or maybe drill presses or steam machines, all communicating with other machines, with control information going back and forth. Those machine-to-machine capabilities were about control, with perhaps a little feedback.

The Internet of Things is really about an ecosystem of communication -- it could be anything from simple command-and-control feedback, all the way to very sophisticated ecosystems.

By its very name, the Internet of Things implies connectivity. To continue with the industrial analogy, we might have a number of devices on factory floors that control different kinds of machinery and that also are extended out to heating, ventilation, and air conditioning controllers. Take those controllers and connect them to the Internet so the signals and data are going somewhere.

We could be talking about something as simple as a cloud-based control system, in which someone is monitoring multiple facilities -- for example, all the Home Depots in North America -- from a centralized facility. All that information comes back there; they are controlling the climate of the buildings from one place. Maybe the data is going off into the cloud, where it's shared with other parts of the organization or partners for supply chain management, with analytics and reporting and control capabilities.

What are some examples of things that can be connected? Are there specific industries that have jumped on this very quickly?

Healthcare is very big in IoT right now. As just one example, you may have heard about a Dell and University of Iowa hospital system case involving surgeons, staff, and the university. When patients are brought into surgical theaters for major surgery, the surgical centers are measuring patient vital signs during surgery, feeding that data into a predictive model in real time, and using the results to predict sepsis -- that is, the probability that a patient on the table right now being operated on will develop an infection after the procedure.

There's also a fair amount of interest in this from retailers, as in the example I gave earlier. They can look at heating and lighting control data from their facilities, as well as different environmental factors.

Also, the industrial Internet is huge in this area, so those are three areas that come to mind.

Regarding the structure of the IoT, what does a typical IoT network look like? How does cloud computing tie into it?

Traditionally and typically, people think about sensors as lying out at the edge of the network. It might be a sensor that listens and watches a lightbulb, or that monitors the number of times a door is opened. ...The sensors are listening to many different things, taking signals from different objects and entities, and bringing the data back into the network. Maybe a thousand different sensors are connected to each gateway. The gateways, in turn, are devices with memory and computing power and different kinds of communications, whether wired or wireless. The gateways take the signals and pass them up to servers; the servers in turn are connected to enterprise data centers or to cloud computing environments.

There's one very interesting aspect to all of this. I just described a single thread in an IoT environment, but the way I think about it is that IoT networks are like circles. The circles are like overlapping Venn diagrams. The center of my network could be the edge of your network. These IoT environments are interconnected networks and the centers and the edges are scattered all over the place and overlapping. That sense of overlapping circles and networks is really key to understanding the IoT and its potential.

Are there some challenges for enterprises when we talk about the IoT?

Privacy is something that is clearly topical, so let's address that. The interconnectivity of these devices is something that is radically different from what has gone before; we talked about that earlier in this conversation. First, consider that many of the signals coming off these interconnected devices inherently have very little to no value. If your sensor is watching a light bulb and it's on or off, it's very unlikely that someone is going to care about the privacy of that signal. However, [we should keep in mind that] even some of the most mundane and simple pieces of information need to be secured and used in a way that ensures privacy.

As we discussed in the recent TDWI Webinar with [TDWI's director of advanced analytics] Fern Halper, the processing of data and the encryption and security of the data can be greatly enhanced if you are processing the data locally. What I mean by that is that you can take the signals that are produced by this factory light bulb and process them and look at them in different ways. You may never have to send the signal back up the network. You can make decisions and filter the information right there. That can greatly enhance the security and privacy of the information because it doesn't traverse back up the network. The locality of analytics, the locality of use, is one way privacy can be greatly enhanced and improved.

That sounds like it takes someone with special skills. Are there enough people out there with the skills to understand IoT challenges and implement the technology?

I think so. Of course, we're always going to be faced with a shortage of skilled people. We've all heard of the shortage of data scientists around the world.

With the IoT, I think it really comes down to breaking things into manageable chunks. When we stop and look at the entire world of IoT and all the various applications and the incredible breadth of applications that we can build with this data, it is daunting. Once we break it down, though, into business objectives and ways to achieve those business objectives, that's much more manageable and easy to understand and achieve.

Much of what we're discussing here is actually common sense. We're interested in using this information and this technology to improve process efficiency, right? Most of us who are numerically literate can understand how operations run -- we can look at our processes and say, how do we improve operations? Maybe we need to focus on energy efficiency. Maybe we want to focus on keeping production runs on time because every time you interrupt a process flow you have waste and loss.

You've been in this industry quite a while. Where do you see us headed with the IoT?

We are moving toward a tremendously exciting future. We have data flows that allow us to do things -- from the most altruistic to the most practical -- that pundits are already writing and talking about online and in the mainstream press.

Here's one great example: there's a tremendous need to make the world greener and I think IoT is going to be one of the great drivers of that. Really, at this point, our operational systems are incredibly inefficient. We're very ineffective in how we use energy and very wasteful in how we design processes and production. I think data and the IoT is going to be one of the ways that is going to help us get to a greener world.

It's also going to help us get to a smarter world. It's going to help us get to one, in my opinion, that will allow us to run things in a much smoother, more automated way. Simply put, we can make the buses and trains run better and more efficiently, we can give riders more information, we can create buildings that are more effective and greener. I see the IoT future as improving the way we live our lives. It will allow us to focus on higher values, on more interesting things. I think the IoT is going to create some truly great opportunities.