Big Data

Embedding Intelligence in the Internet of Things

Share this post:

Three years ago when my wife and I were visiting Paris, we happened upon the finish of the Tour de France bicycle race. We were swept up in a huge crowd, and we were thrilled to watch bunches of riders zip around the Champs-Elysees at blazing speed. Amazing!

The only problem was, we had no idea what we were seeing.

Fast forward to last year, and it was a totally different story for people watching the race anywhere along the route. From their smart phones, they could follow in real time the progress of the ultimate winner, Britain’s Chris Froome, or any other rider of their choosing. People who did not attend could tap into all sorts of stats on the Tour Website—enabling them to do their own analysis of the race. Meanwhile, TV announcers could report the exact location, speed and competitive groupings for each of the 198 racers at any moment during the 21-day event—greatly enriching their color commentary. It was a huge breakthrough for all involved.

In the past, the race organizers gathered barebones stats from staff members on bikes who were equipped with stop watches—an arduous process that delivered spotty and much-delayed results.

What changed things was putting GPS sensors on every bicycle and installing a sophisticated relay system that transmitted data to central servers, and then to apps, Web sites and broadcasters. The racers and their bikes were data-collection points on the Internet of Things. The goal was nothing less than transforming the Tour fan experience.

One of the key ingredients in this setup was IBM Streams technology, which gathers data from many sources in real time and organizes it so insights can be drawn from it. In the future, thanks to a new but related IBM technology, Quarks, cycling fans, broadcasters and team strategists could gain much deeper insights during the Tour. Quarks inserts intelligence into the Internet of Things.

If you want to explore Quarks, check out activities on GitHub

We named Quarks after the physics term, quark, which is an elemental particle and a fundamental constituent of matter. In tech parlance, Quarks is a “runtime” version of the Streams technology that’s stripped down so it can function as a tiny digital nervous system. In the bicycle racing scenario, Quarks has the potential to be combined with a wide variety of IoT sensors, including GPS, gyroscopes, accelerometers, barometric pressure sensors, etc. A reading from a gyroscope could tell you if a rider has fallen. You could attach health sensors to a rider and feed data to his race director—so they can make better decisions based on physiological data and power output.

The ability to pack a wide variety of sensors on a device has its upside and its downside. The positive is that you have an immense amount of data. The negative is that you have an immense amount of data. By embedding Quarks in the devices, you can perform some of the data management and analysis right there, on the edge of the network, and send only the key bits of data to your central computing systems. In essence, you’re generating insights at the data source.

The Quarks technology, in combination with Streams, has the potential to transform whole industries, including health care, manufacturing, mining, transportation, oil & gas, electric utilities, and automobiles; and to improve the management of cities, highway systems, public safety, and storms. We anticipate connecting a wide variety Internet-of-Things devices to our Watson cognitive computing systems—so the IoT becomes a massive feeder of all types of data to Watson.

We invented Streams more than 10 years ago to manage data in motion—which is essential for producing actionable insights in real time. We tested it in a variety of situations, including on the premie ward of a major hospital and on a securities trading floor. Then we branched out to a wide array of uses, ranging from water management to radio astronomy particle physics.

Last year, we began to get feedback from our clients that they wanted to include intelligence in their Internet-of-Things sensors. I had a series of conversations with Mike Spicer, the lead architect for Streams, and, together, we reached the conclusion that the time for an embedded version of Streams had come. Thus, Quarks was born. We decided to make Quarks an open source project so all sorts of people and companies would feel comfortable putting it in their products—hopefully, accelerating adoption.

We believe that a lot of our clients will use Quarks in conjunction with Streams, but that’s not necessary. In fact, we can envision scenarios where devices outfitted with Quarks will communicate directly with one another. Cars on a highway, for instance, might warn other drivers nearby if there are icy conditions ahead.

In recent weeks, we’ve been approaching clients to gauge their interest in Quarks, and we’ve received positive reactions. For example, Cummins Engine. It’s the largest maker of industrial and truck engines in the world, but increasingly makes money by providing aftermarket services for its customers. It installs sensors in its equipment to monitor operations and manage maintenance. The company was already using Streams. Last fall, Cummins said they wanted a standardized way of embedding intelligence in the sensors, no matter who supplied the individual devices, so they could perform on-vehicle analysis of data. Quarks fit the bill.

So you can see the potential of Quarks. It’s a big deal for everybody from bicycle racers to truck drivers. But I think the biggest impact will be in healthcare, where imbedding intelligence in medical monitoring devices will enable sick or elderly people to live at home rather than in the hospital, and early alerts about dangerous medical conditions will save countless lives.

This is the kind of impact we dreamed of when we invented Streams. It’s intensely gratifying that our dreams are now coming true.

More stories

Accelerating Digital Transformation with DataOps

Across an array of use cases, AI pioneers are employing a core set of new AI capabilities to unlock the value of data in new ways. According to the 2019 IBM Global C-suite study, leaders are using data 154% more to identify unmet customer needs, enter new markets, and develop new business models. These leaders […]

Continue reading

The 3 Beachheads of AI

(Part 2 in a Series) We have been partnering with and assisting clients on their data needs and strategies for years. It’s clear that data and AI are two-sides of the same coin; in fact, this understanding spawned the AI Ladder concept. We’ve developed skills training in the areas of data science and machine learning; […]

Continue reading

Medicine and the Message: Consumer Engagement and the Power of Nontraditional Data

The journal Health Affairs estimates that the U.S. spends nearly $500 billion annually on pharmaceutical treatments. Focusing on clinical value alone isn’t enough to meet business growth targets — customer experience plays a critical role in long-term brand success. But experience doesn’t occur in isolation: To win trust and boost spending, pharma brands must use […]

Continue reading