It’s hard not to bump into the Internet of Things (IoT) these days. The last few months alone have seen a flurry of activity. That list includes Google’s $3.2 billion purchase of Learning Thermostat maker Nest in a big play to advance its connected device efforts; as well as, IBM and AT&T joining forces to develop IoT solutions that initially will help city governments improve things like traffic management. Governments are trying to move the ball forward themselves, too. In March, for example, U.K. Prime Minister David Cameron touted an additional £45m in research funding to develop the Internet of Things, bringing the total to £73m, plus another £1m to support startups in the area.
Frankly, it’s impossible to browse the Web without coming across another story of a new gadget built to connect to the Internet of Things, whether it’s a toothbrush or a beer keg. Soon enough, those bumps will become collisions: Cisco has predicted that 25 billion devices will be connected by 2015, and 50 billion by 2020.
What impact will a growing Internet of Things have on the enterprise? Certainly there is opportunity for companies to create new apps for monitoring and controlling this emerging world of connected devices. And the door is wide open for them to develop innovative products and services from their analysis of the data coming in from all the devices connected to and communicating with each other, and with other smart objects, over the Internet.
It’s no surprise that companies will face some trials as they enter this brave new world, as well. Now that it’s possible to wire up sensors on machines, not to mention humans and apps, “to pull all sorts of data points out, the challenge is … whether you're capturing the right information, and how to make decisions [and] extract trends from that data,” says Dan Keldsen, president and principal consultant of Information Architected, which helps companies develop methods to intelligently use their corporate content and knowledge.
In addition to becoming better at leveraging data from IoT devices for insight, businesses have to consider revamping Data Management strategies, and perhaps even their infrastructure choices, to deal with the tidal wave of information coming their way. “If you think about 50 billion devices connected by 2020, imagine them all constantly providing data back into a central hub to massage, manipulate and process,” says Robert E. Stroud, CGEIT, CRISC, member of the Professional Influence/Advocacy Committee at ISACA – an independent, nonprofit, global association focused on the development, adoption and use of globally accepted, industry-leading knowledge and practices for information systems, and vice president of Strategy and Innovation at CA Technologies. “We are going to see an exponential increase in data.”
What To Do With All That Data
Consider, as an example, what things will look like for the typical car manufacturer, Stroud says. Today, many new cars come with sensors that collect and process information about everything from engine mechanics to seatbelt activation within the vehicle itself. But in the age of the Internet of Things, that information can travel the data highway in real-time from every new car that rolls off the line back to a central hub in the manufacturer’s systems, where it can be aggregated and analyzed to support decisions about how to improve vehicle makes and models under various conditions.
Say that a particular car model is experiencing problems skidding on the road. In that case, it could be helpful to match incoming sensor data from those models’ about such events with corresponding data, perhaps from other sources, about the weather conditions at the moment when and in the place where these incidents occurred. “The secret of IoT is that it is great to have things interconnected, but you have to correlate information to make relevant decisions,” Stroud says. “It’s not just a case of logging and storing data forever, but about being diligent about what you use.”
In other words, when it comes to leveraging IoT data, have a valid business problem and outcome in mind from the start, he advises. Don’t just collect data for data’s sake. In the above scenario, for instance, the analysis likely wouldn’t be improved by also incorporating data from the affected vehicles’ ambient light sensors. “If I just put sensors in cars for data for no particular reason or outcome, I’ll just spend a lot of time and money and may not actually get any value out of it.” Stroud says.
While sensor data related to turning a car’s lights on and off might not matter to the current business problem the car company is trying to solve, that doesn’t mean that it can’t be useful to the business in other ways in the future. There may be hidden patterns in the overall IoT data stream of which it is a part that could be useful further down the road. In the future, “a data scientist may want access to that voluminous amount of information for a special project,” Stroud says. “Data archiving might be back in vogue again.”
Companies have to think about how to allocate the right amount of money and resources to make use of historical data vs. new or current data for decisions they’ll be making now and in the near future, adds Keldsen:
“Even with storage costs plummeting, most organizations can't afford, and it really wouldn't make sense, to leave everything in their highest performing storage systems for ‘instant access,’" he says. “This is where the discipline of traditional data warehousing needs to continue to play a role. Being sloppy about storage retention can cost you quite a lot if you aren't careful about balancing out business needs vs. the associated costs.”
Can You Play It Safe?
Companies also likely will come up against security and privacy issues as they dip their toes into the Internet of Things. As consumers adopt wearable devices or other sensor-based gadgets, for instance, they may agree to let companies track them in exchange for certain perks. They could be offered lower insurance premiums if they agree to having a device in their car to monitor their driving patterns. But the duty of care – be it legal or moral – that companies have long had to abide by for respecting and securing personal data feeds could become even more exacting.
Stroud believes that the IoT will give birth to an industry of third-party providers who will collect, collate, and distribute sensor-based data. One outcome of that is that the information consumers explicitly agree to share with a company may get merged with IoT data that the same company may secure from a third party. The result – however unintentional – may be that the company gets a more complete picture of the user than he is aware is being taken, or would be comfortable with if he did know:
“We haven’t thought through all the ethical ramifications yet,” Stroud says. “We need to manage the hygiene aspect of data like we would normally, but the reality is the scale of this effort will dramatically increase with the additional information.”
Another caution when it comes to basing decisions on information gleaned from the Internet of Things is that conclusions could be inaccurate if the company arrives at them without having enough data points, or without having the appropriate ones. “This is still a relatively new science so you almost have to double-check results before doing something with your information,” Stroud says.
While there’s no reason to hold back on moving in the direction of the Internet of Things, just understand that there is still some risk around outcomes, he says. The good news is that things will get better “for sure,” Stroud believes. Not only are a growing number of individuals interested in pursuing Data Ccience as a profession, but there already are people in organizations who have the right skill sets. Says Stroud, “Their experience and how they approach decision-making are what we need to emulate in managing information from the Internet of Things.”