In the United States, around 200,000 manned U.S. general aviation aircraft have been registered over the last 50 years. By contrast, 750,000 unmanned aircraft systems — aka drones — have now been registered, including more than 40,000 in the last two weeks of December 2016 alone. And that’s just one country. It exemplifies the dramatic influx of “things,” which carries unprecedented opportunity for digital disruption. They’re typically full of sensors, increasingly connected, produce enormous amounts of data and can be the source of newer, smarter business models that touch every industry. For example, in the past decade, wind turbines have quickly evolved from isolated standalone machines to connected, sensor-laden, intelligent devices. One of the largest suppliers of wind energy, Vestas, has 60,000 turbines, and the newest ones have 1,000 embedded sensors each, all capable of emitting real-time streams of data. How can we harness the power of such massive amounts of real-time, streaming sensor information? How do we manage computation at drone scale?
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
Drone-scale computing requires new thinking. Over a decade ago, sensor-based computing research began through funding by the United States intelligence agencies at MIT; in parallel, academics at Stanford and Cambridge also explored how sensors would change the computing physics of moving data. That academic research spawned commercial and open source stream computing technologies, and some have evolved into streaming AI engines. Just this past year, Forrester included streaming analytics as part of its next-generation business intelligence category, which it calls “systems of insight.” While last-generation business intelligence is focused on putting data on a graph, systems of insight are focused on generating insight and directing action — and are being deployed for transportation and logistics systems, digital customer engagement and intelligent industrial IoT applications.
A system of insight is like your human nervous system: AI is the brain, IoT sensors are your senses, middleware is your skeletal system and streaming analytics complete the autonomous nervous system’s function.
The brain is fueled by sensory input. In a system of insight, IoT sensors are like nerve endings in your fingertips. Signals are captured and distributed throughout the nervous system. But our IoT sensor nervous system is evolving. Today, many IoT applications must rely on decades-old SCADA networks. Thanks to drone-scale requirements and improving networks, that archaic IoT fabric is being replaced by a new infrastructure. AI is now used to produce algorithms, which must be injected into the network nervous system to enhance capability. Firms like GE, Siemens, Rolls-Royce, Syniverse and QIO are all in a sensor network arms race to create smarter IoT networks that can extract and transmit data from sensors in real time. The Vestas implementation captures terabytes of sensory input each and every day from its wind turbines to continuously train algorithms that continuously instruct turbines on how to react to wind and atmospheric conditions and optimize power production. It considers that system of insight one of its most valuable — and secret — corporate assets.
Streaming analytics is analogous to your autonomic nervous system. You don’t think about taking your hand off a hot stove, you just do it. Athletes rely on their nervous system to strike a golf ball, fake out a defender or pass to a teammate behind the back. Similarly, drone-scale systems require autonomic reaction to conditions. Streaming analytics of trained algorithms provide this automatic intelligence in action for IoT systems and guide delivery drones to, say, avoid collision with one another. It doesn’t have to “think” about it, it just has to move. Such prowess is achieved by network operators and businesses combining multiple sources of streaming data — for example, correlating mobile transactions with drone movements and delivery. This creates real-time “game awareness” like that of great athletes and coaches who see the whole picture and make smart choices: Do I run a different play, choose a different receiver or call a timeout? Algorithmic muscle memory and algorithmic game awareness depend on correlated streaming AI analytical models.
Our body’s skeletal system differs somewhat from IoT skeletal systems — for one, our bodies don’t have to worry about hackers and security breaches (yet!). All automated IoT systems face this challenge, as the importance of a connected business nervous system makes for a big target. IoT systems must be as secure as they are smart, from the edge (on a device) to the core (for network-wide visibility). To function at drone-scale, security must be algorithmic and automated, and incredibly diligent as well. In 2012, the automated trading system for Knight Capital lost $440 million in 40 minutes due to an automation flaw. Algorithmic security is one of the most important new fields of our IoT-powered future.
The final question is, with all this intelligence and automation built into the new drone-scale nervous system, are humans still needed? The answer is absolutely yes! But the role of the human shifts. While repetitious, manual tasks become more automated, new, even smarter roles for humans emerge. For example, on Wall Street, where 80% of the world’s trading activity is algorithmic, humans are more essential than ever. The trading pits of the ’90s may be gone, but now humans play higher-value customer relationship roles — even average-sized trading firms manage billions of events a day, so algorithms augment the ability of human staff to decide who to help, who to upsell and who is at risk. Those customer advisory algorithms, like the trading automation itself, is one reason for the dramatic rise in demand for data scientists. So, in a drone-powered world, the role of humans will be more important than ever, but the job descriptions will change to best utilize the new tooling.
The next generation of IoT-powered business will emulate the human nervous system from the edge to the core of the IoT fabric in the cloud. The opportunities to leverage this nervous system are endless across all industries. “Things” that were once isolated — from your coffeemaker to wind turbines to drones — now represent the biggest disruption opportunity of the future. We are entering a world where automated, AI-driven action happens faster than the blink of an eye, embedded in streaming IoT sensor data, and aflight all around us. Are you cleared for takeoff?
All IoT Agenda network contributors are responsible for the content and accuracy of their posts. Opinions are of the writers and do not necessarily convey the thoughts of IoT Agenda.