2015 Top Embedded Innovators: Dr. Stan Schneider, Chief Executive Officer, Real-Time Innovations (RTI)
Elected by his peers, Schneider serves on the Industrial Internet Consortium (IIC) Steering Committee. The role includes fiscal oversight and setting up subcommittees such as research selection, test bed creation, and defining standards requirements and priorities. In addition, Schneider has an impressive record of academic and industry published papers and speaking engagements on topics ranging from networked medical devices for patient safety, the future of connected cars, the role of the DDS standard in the IIoT, the evolution of power systems, and understanding the various IoT protocols. Embedded Computing Design asked Schneider about addressing the challenges to innovation, how to be an innovator and recognize areas for potential innovation, and what he sees as the next big market and technologies for embedded computing.
What are the largest obstacles to innovation in the embedded space, and how should those challenges be solved?
The "Internet of Things," or IoT will be the next great wave of connectivity after the Internet and the mobile revolution. The IoT, however, is bigger than both combined; there will soon be billions upon billions of devices. This will be the golden age of embedded.
Wearables, home thermostats, and other smart consumer devices will become part of daily life. That said, the big economic impact is not from smarter consumer gadgets. The real value is the opportunity to redesign our most critical infrastructure: transportation, energy, medicine, and manufacturing. These "Industrial IoT" distributed systems can get smarter and be more efficient, safer, and faster. Entire new classes of systems, like autonomous vehicles, smart, connected hospital rooms, and high-performance "microgrids" will fundamentally rewrite the way our world works.
The largest obstacle to this vision is the lack of "systems thinking." Traditional embedded systems, and consumer IoT for the most part, are standalone systems. Designers think of what chip, OS, and perhaps wireless or network connectivity they will need for their "device." This is "device" thinking.
The Industrial IoT, on the other hand, is a "systems" problem. A connected IIoT system will run on many operating systems, networks, and devices. Those become secondary concerns. The systems problem is not about devices, it's about their interactions. The challenges are interoperability, overall systems data management, discovery of data producers and consumers, and scale. There is no way to build these systems with traditional programming techniques. Object orientation, for instance, works well for contained programs. But, exporting "methods" that can vary across a huge system means that every interaction is custom and every interaction is one-to-one. That's unmanageable at scale. Big systems are all about the data; the Industrial IoT embedded systems of the future will need a data-centric interaction approach. Data-centric interaction means modeling the data itself. Data-centric interfaces control data interactions like discovery, rates, reliability, delays, and security.
How do you stay on the leading edge of innovation, rather than just following the embedded crowd?
Well, problem number 1; the embedded crowd thinks of themselves as the embedded crowd. I don't mean to be cute, but the way to innovate today is to apply "Internet thinking" to everything. That is the power of the IoT.
One of our customers, BK Medical, is my favorite example. BK makes ultrasound scanners. Today's ultrasound machines have displays, interchangeable transducers, and image processing hardware or software. They produce physical media, usually a DVD. They roll into the clinic on wheels. The problem? This fundamental "embedded system" looks the same as it did 30 years ago. Back then, it would have a CRT instead of a flat screen and you'd get a Polaroid instead of a DVD. But, the architecture is the same. It's a standalone embedded system.
BK's new design will instead give every doctor a wireless transducer to carry in his or her pocket. The raw data will go over a network to the cloud (hosted in the hospital for now) for analysis and image production. The result will be immediately displayable, put into the medical record, and even billed to the patient. There's no need for physical media because the image is already online. BK is replacing the standalone embedded system on a wheeled cart with a fully distributed system. It gives every doctor exactly the right transducer. It eliminates searching or waiting for the machine. And, it saves the hospital space, time, and money.
This is the destiny of embedded. "Embedded systems" used to mean a single processor in a single device. In the future, embedded systems will truly be systems: distributed smart machines encompassing dozens, hundreds, or millions of processors working together as a single intelligent whole.
How do you recognize when a new technology or application is one your company should invest/innovate in, versus a technology that will experience fast burnout?
For us, it's simple. We are laser-focused on systems that combine needs for mission-critical reliability (can't go down for 5 seconds, or even 5 ms), performance or scale (measured in ms, or 10k data items), and an infrastructure life cycle (last more than 3 years). These large systems are the burgeoning infrastructure of the next generation of the industrial revolution.
More generally, we look for leverage due to connectivity. What can a distributed, intelligent system do that's really new and valuable? Another example from the medical industry: hospitals are packed with hundreds of types of devices, from respirators to oximeters to ECG monitors. Today, they are standalone systems. Working alone, devices find worrisome readings for all sorts of reasons. The resulting nuisance alarms are a real problem. As a result, 80 percent of all alarms are turned off. There are whole conferences on alarm fatigue. Patients go unmonitored. And, tens of thousands of people die every year from preventable errors that smarter alarms and smarter distributed connected device systems could easily prevent.
This is not just medical; poorly connected infrastructure is the rule, not the exception. Much, if not most, of "green" energy generated by solar cells is wasted because the power grid is too poorly connected to save shut down "spinning reserve" backup generators. We spend untold billions on carpool lanes that are only valuable because we cannot properly control lanes to keep freeways moving. Factories and wind turbines and imaging systems and trains fail because we don't monitor their state and predict failures. These optimizations are critical beneficiaries of the Industrial IoT. We (RTI) see innovation in providing the connectivity that enables solutions. Others find innovation in using that newfound capability to build analytics and applications and services that can use the newly available data.
In the next 5 years, which embedded technologies, applications, markets, and geographic areas present the most interesting opportunities?
The next 5 years will be the most transformational in embedded history. Every market, from space to mining, from transportation to hospitals, from power to manufacturing faces the opportunity to distribute, connect, and transform. We live in a magic age.