Internet of Things applications put varied machine data to work

The Internet of Things is proving to be a hotbed of diverse data and new applications fueled by information from sensors and other machine sources.

In the pantheon of recent buzzwords, the Internet of Things (IoT) has garnered a special place. It isn't hard to see why: "Things" covers a lot of ground, doesn't it? The buzz may be off the term by the time we see a lot of IoT applications -- the road may be longer than the IoT's most boisterous advocates imagine. Still, you can already find some interesting application types, ones that involve new ways of working with data.

Down in the Flint River Valley of Georgia, there are a lot of farms producing a lot of peanuts, cotton and corn. Those crops are nourished by irrigation systems that could become more efficient with the application of analytical power. At least that's the estimation of Lloyd Treinish, chief scientist for IBM's Deep Thunder weather analytics and forecasting service.

He and IBM colleagues are working together with the government-funded Flint River Partnership to integrate the supercomputer-based Deep Thunder system with data feeds generated by in-field weather stations and soil moisture sensors as well as satellites, commercial weather data services, GPS-enabled farm equipment and field sensors.

The Flint River Partnership includes the Flint River Soil and Water Conservation District, the U.S. Department of Agriculture's Natural Resources Conservation Service and The Nature Conservancy. One of its goals for the analytics system is to help tune irrigation schedules for the valley's crops based on immediate weather information -- that is, 72-hour reports updated every 12 hours, forecasting conditions at 10-minute intervals in 1.5-square-kilometer areas.

Water, water everywhere -- but not too much

You can't do much about changing the weather, but you can do something about the amount of irrigation that crops require, Treinish told us. The problem now, he contended, is that conventional forecasts don't provide enough information to be helpful there.

"For planning irrigation, today's weather forecasts are almost useless," Treinish said. "They're vague: It's something like 'chance of thunderstorms in the afternoon.'" But, with a variety of data feeds and the right algorithms, he hopes to help farmers in Georgia -- and elsewhere, eventually -- achieve highly efficient irrigation scheduling that provides both economic and environmental benefits.

The precision that Treinish and Flint River Partnership officials are looking for is daunting. The system is meant to tell which small acres of large tracts will be inundated with rainfall and which will merely experience light rain. Ambitious? Yes. But some grand ambition is what many people have in mind when they point toward the IoT.

The ongoing work in the Flint River Valley gives one view on how IoT applications may come to be. Another comes from a conversation with Fred Gordie, a technology evangelist at McKenney's Inc., a building design, construction and maintenance services company headquartered in Atlanta, Georgia.

How green was my building?

Software has become an ever greater part of what a building services company like McKenney's does. Gordie said McKenney's uses Splunk Enterprise as an analytics data integration platform for a business intelligence system it offers for analyzing facility operations. One showcase for the effort is Eglin Air Force Base, a sprawling complex that occupies 724 square miles of land in Florida and contains more than 2,000 buildings. McKenney's has been a part of an effort to analyze machine data from HVAC systems, elevators and more as part of a multi-building energy management system.

In recent years, McKenney's has become good at gathering data by what some might call the IoT, but the company looked to improve its analytical prowess as well in order to take advantage of all the data, according to Gordie. "We had a control system and power monitoring system with sensors," he said. "The task was to bring all the information together and to quickly and dynamically do data analytics on it."

Building sensors produce a wide variety of data, creating analytics challenges, Gordie said. In fact, McKenney's tried using a SQL database to run its bdocs BI system but found it was a difficult task to combine the varied types of machine data in a timely and flexible fashion. The development team eventually opted for the Splunk technology, which has seen some traction in the machine data arena.

Check out recent Talking Data podcasts and columns

Learn about the clash of cloud and data

View the latest take on big data trends

Discover the potential of cognitive computing

Gordie said the developers were able to create data relationships flexibly, "on the fly," with Splunk's index-oriented approach. "We've yet to find data that Splunk can't eat," he noted. And being something of a data omnivore is likely to be good thing in the new world of IoT applications.

Those are just a couple of snapshots on a random walk around the IoT, but surely there are a lot of "things" out there available to be connected on the Internet. The IoT is potentially pretty vast, even bigger than the World Wide Web. The vastness makes it a magnet for hyperbole -- but it seems possible to uncover good uses for the new data technologies. The feeling here is that there are lots of signals in the noise of all that buzz.

Jack Vaughan is SearchDataManagement's news and site editor. Email him at [email protected], and follow us on Twitter: @sDataManagement.

Next Steps

Learn how to speed up building your IoT application for the cloud


Dig Deeper on IoT APIs, Applications and Software

Data Center
Data Management