Merger of Software and Hardware Has Big Implications for Manufacturing and Design

A Makani airborne wind turbine, a lightweight wing that's said to eliminate 90 percent of the material used in conventional wind systems, was on display at O'Reilly Media's first Solid Conference in San Francisco in May.
Photo courtesy of O'Reilly Media, Inc. (

Mark Shortt
Editorial Director
Design-2-Part Magazine

The impact of bold new interactions between software and hardware—from connected cars to drones, flying wind turbines, and even runaway smart balloons—is just beginning to be felt.

Remember when Google was known only for being a search engine? A dominant one, of course, but just a few short years ago, there weren't a whole lot of other ways to describe what the company offered.

That's no longer the case. Within the last year and a half, Google has been on a buying spree, snapping up hardware companies in the robotics, machine vision, and home automation sectors. Companies like Nest Labs, Boston Dynamics, Redwood Robotics, Industrial Perception, Makani Power, and Titan Aerospace are just a few of the 40 companies—including eight robotics manufacturers—acquired by Google since the beginning of 2013. What's Google trying to accomplish by picking up companies that actually make things? While company executives remain tight-lipped, the acquisitions are perhaps the most visible indicators of a nascent, rapidly accelerating global trend toward building an Internet of Things (IoT), a merging of the digital and physical worlds that figures to have profound implications for the design and manufacture of components, devices, products, and equipment.

Tom Racciatti, a senior manager in the mergers and acquisitions practice of West Monroe Partners, a management and technology consulting firm, has been keeping a close eye on mergers and acquisitions involving prominent tech companies and manufacturers. In a recent blog post, Racciatti said that in addition to Google's acquisitions, the company's significant investments in the development of Google Glass, Google Fiber, and self-driving vehicles point to a "direct, explicit strategy to move into hard goods." To Racciatti, the interest that Google and other tech giants—such as Samsung and Cisco—are showing in the manufacturing sector makes sense. The new wave of Internet connected products, he believes, represents an opportunity for tech (software) companies to create a new market for their current tech platforms.

"I think that while maybe they're not being as explicit as it appears to the rest of us, it seems apparent that they're clearly trying to play in to the Internet of Things, hoping that it is the next huge market dynamic that they can capitalize on early," said Racciatti in a follow-up phone interview. "I don't think they want to be the world's best manufacturer of the actual piece of hardware. When it's Nest, I don't think they care what the thermostat or even the features on your phone look like. But it's the size of that prize of connecting everything. I think they [Google] recognize, and a few other companies are recognizing, that the size of that prize is massive."

Cisco reported earlier this year that, based upon its own research, the company estimates that more than 50 billion devices will be connected to the Internet by 2020, creating what company representatives say are "vast opportunities for the Internet of Things to transform businesses and industries by re-inventing business processes, operational efficiencies, and customer service innovations." According to a release from Cisco, these opportunities are waiting to be tapped in numerous markets, including manufacturing, transportation, oil and gas, utilities, healthcare, sports and entertainment, government, education, and retail.

Google and other tech companies' moves into the hardware sector are likely to create new opportunities for U.S. manufacturing companies—especially contract manufacturers—who can help them bring their new products to market. There's room for growth on both sides—for tech firms and for contract manufacturing companies—as the need for mutually beneficial partnerships increases.

"I think it's good for them (contract manufacturers)," said Racciatti. "As I alluded to earlier, I don't think Google, for example, has any interest in making, in actually becoming the best manufacturer of their hard product. I don't think there's any likelihood of their moving that far from their sort of core competency. I think what that means is more and more opportunity for contract manufacturers to not only, essentially, have new volume and create products for new businesses, but also to get more sophisticated by working with those software companies."

Meanwhile, it appears that both Google and Facebook are looking for a little help from high-altitude drones, or unmanned aircraft, to make Internet access a reality for more people across the globe. In April, Google bought New Mexico-based Titan Aerospace, a company that Facebook had been in talks to buy, and Facebook acquired the U.K.-based aerospace company, Ascenta, in March. Google is also working on Project Loon, a radical attempt to bring Internet access to remote, underserved regions of the world through the use of smart, helium balloons.

Will tech companies' appetite for hardware manufacturers continue? Google has hinted that more acquisitions will follow, and Racciatti said he doesn't see any imminent roadblocks to overall activity within the sectors. "I think that continues—until when, I don't know, but I do see what we've seen recently continuing," he said. "That is, more and more acquisitions and more and more of big technology companies moving in where they can to try to get the Internet of Things prize."

The Internet of Things "prize" is big—some analysts project the market to be worth close to $500 billion by 2019—and there appears to be no limit to the markets that IoT will penetrate. Home energy management, home appliances, automotive, aerospace, energy, medical devices, and industrial automation and manufacturing are some of the biggest markets for connected products. Current efforts to apply IoT to manufacturing include the Industrial Internet Consortium (backed by GE, Cisco, IBM, and other leading companies) and the Smart Manufacturing Leadership Coalition (see Laying the Groundwork for Factories of the Future).

The Wall between Software and Hardware is Crumbling

Something new, intriguing, and really big is brewing inside Fort Mason Center's Festival Pavilion, a renovated military building on a narrow peninsula poking out into San Francisco Bay. That much is clear to the estimated 1,000-plus technophiles who've already made their way through the doors of the Pavilion on this cool, breezy Thursday morning in late May. They've come to attend O'Reilly Media's first-ever Solid Conference (SolidCon 2014), a two-day event that's bringing people together from multiple disciplines—business leaders, software developers, hardware and mechanical engineers, product managers, and marketing execs—to explore what is now widely anticipated to be the far-reaching convergence of software and hardware and what this might mean, even beyond the development of intelligent new consumer products.

Prototype engineer Alex Wickersham installs a weight to ensure that Makani's newest Wing 7 flies evenly.
Photo courtesy of Makani/Google.

That convergence, reflected in the conference theme "Software/Hardware/Everywhere," is being explored through presentations like "The Future of How Things Are Made," "Integrated Product Development at the Intersection of Hardware and Software," "The Industrial Internet and Manufacturing's Digital Revolution" and "Smart Privacy: Rules of the Road and the Challenges Ahead." At the same time, a variety of slick demos and exhibits of products—ranging from an airborne wind turbine to a high-definition 3D printing machine to a wall-climbing robot—provide tangible examples of a new wave of product development.

One of the exhibitors is Jamel Seagraves, a research engineer at Ford Motor Company's Silicon Valley Lab (Ford SVL) in Palo Alto, California. A tall, affable man, Seagraves is explaining Ford's open source platform, known as OpenXCTM ( It's a research platform that combines open source hardware and software to enable users to extend their vehicles with custom applications and pluggable hardware modules, according to

"Its primary focus is for hardware and software applications, sort of centered around 'How can we extend the vehicle and make it more of a connected car?' Seagraves tells D2P. "This is really kind of Ford's first step in tapping into and reaching out to the maker and developer communities to try to spark that idea generation and brainstorming, that tinkering and development that maybe creates ideas that we wouldn't have thought of, internally, on our own."

The primary way that they're able to do that, he explains, is by opening up CAN (controller area network) signals to give the user real time, live data from the car. "We have a vehicle interface, which is essentially a dongle that plugs into the OBD-II (on-board diagnostics) port. "As a car company, we're able to give you that real time data, and so we've currently released 20 signals to the public, open source. Even the design of the dongle itself is open source.

"So it's really just trying to spark kind of that inspiration and see what we can make of it. It's not necessarily meant, at this time, to be a Ford product that we try to sell or push on people. It's more of "What creative things can you do with the data that comes with it? Maybe you don't want to do anything with the data, or maybe you want just basic things done with it."

Ford SVL has held some hackathons, Seagraves said, to see what kinds of things people might come up with. So far, the hackathons have yielded a number of interesting software applications in areas like traffic management, monitoring of electric vehicle battery status, and using location data from the car's GPS to create parking tools. At this point, Seagraves said, he and his colleagues at Ford's 12-person Silicon Valley Lab have no particular preference for what they're hoping to get out of the open source project.

"I think it's more of an exploration because, really, the users are going to have more of that vision—even more, maybe, than we would have on our own if we tried to push that on people," he explained. "And so we kind of just want to let them work in an open sandbox and really see what they can come up with. Maybe somebody's interested in the safety aspect and they come up with something great; maybe it's more on the side of mobile connection, right, or connected to your house, or something, in that way. So we want to leave the sandbox as open as possible and capture as much as possible."

Big Data in the Sky

Another exhibitor is Pramila Mullan, senior manager at Accenture Technology Labs, which has been actively exploring and testing the use of unmanned aerial vehicles (UAVs), or drones, in recent months. Ms. Mullan, who's scheduled to speak in a presentation later that afternoon on the role of unmanned aerial vehicles (UAVs) in oil and gas pipeline maintenance, tells D2P that UAVs are a relatively new initiative for Accenture.

"We've been in the Industrial Internet for a very long time, but the work around the use of UAVs has really been within the last year," she said, adding that there are essentially three factors that are driving the demand for the technology in two key fields of application—pipeline monitoring and inspection for the oil and gas industry, and agriculture.

"There are particular reasons why we are gravitating towards those vertical markets and why we feel they will be the early adopters. One is certainly safety—inherent safety. The value of human life is irreplaceable, so it's really from that perspective," she said, alluding to the risks involved in monitoring pipeline production platforms in areas that are typically remote and often dangerous, such as deep water and arctic environments. "Another factor is cost. UAVs give you the ability to do certain monitoring in a much more reliable way and more cheaply than other options, especially for oil and gas. The other factor has to do with privacy considerations in more urban areas."

Mullan said that although there's been a lot of hype around the use of UAVs to do delivery of various types, "whether they are medicines, or packages, or tacos or burritos," adoption of the technology will likely proceed more slowly in urban settings due to concerns over privacy. "It's going to be a little bit easier to get that adoption going in non-urban settings, like, for example, in farms, where you've got private ownership, or in oil and gas pipelines, which are typically in remote areas," she said. "There are no privacy concerns, except for those of the pipeline operators, and the data is for their own use."

When asked if there were any particular design challenges associated with UAVs, Mullan said that Accenture Technology Labs works closely with its partners—many of which are large enterprises and industrial firms—in developing their roadmaps from a design perspective. "We have that constant feedback loop," she noted. "This is an ecosystem that is collaborating together, and we're very much about bringing that ecosystem together so that we can learn from each other in order to improve the overall, end-to-end solution. Battery utilization comes to mind; the other thing that comes to mind has to do with hardening the UAVs for different types of terrain and weather conditions. But having said that, this type of technology came out of the military, and the military really were leading edge in solving some of these problems. It's just a matter of time before we find them coming into the commercial mainstream." For more on UAVs, see North Carolina: First in Flight Once Again?.

In a separate sit-down interview with O'Reilly Media's Editor-at-Large, Jon Bruner, Mullan said that although drones represent a great opportunity for many of Accenture's clients to achieve cost savings, a lot of work needs to be done around the use of the technology, particularly with regard to the data that it makes available.

"It's not about just one drone or having five drones; it's about having a fleet of drones and then, more importantly, what do you do with that data once you get that data?" she said in a video of the interview posted on YouTube ( "So this is where Accenture comes in. We have a lot of strengths in integration; we have a lot of strengths in data and analytics and data insights, and developing dashboards that allow us to understand and get some meaning out of that data. And so what we're essentially doing is we're taking these UAVs, we're using them as a platform to stream real time videos' data. Also, we're using it as a platform for collection of sensor data that can potentially be embedded in the ground, and then processing that through our analytics engines and then taking them on."

Mullan also told D2P that Accenture's work in the broader field of unmanned operations extends beyond UAVs. "We're also looking at collaborative robots, and how they change the manufacturing floor, in order to be able to leverage more of that technology to do bigger and better things for plant operations," she said.

What Google X Sees in Hardware

The opening day of the Solid Conference featured a keynote presentation by Google X's "captain of moonshots," Dr. Astro Teller, who laid out in spellbinding fashion why Google X is embracing the challenges of working with hardware, and how that choice has shaped the company's culture and engineering efforts.

In a video posted on O'Reilly's, Teller is seen striding firmly back and forth across the stage, illuminating his words with humor, long sweeps of his hand, and a tone that mixes wide-eyed wonder with the sharpest of analytical insight. The man who oversees day-to-day operations of Google X, Google's "moonshot factory," Teller is explaining what's behind Google X's focus on hardware—described as a "focus on atoms, not just bits"—in his keynote, "Google X's Focus on the Physical World." He begins by sharing an anecdote from an experience involving Project Loon, a radical attempt by Google X to bring Internet access to remote, underserved regions of the world through the use of smart, helium balloons floating in the stratosphere.

"All teams are used to tracking the progress of their projects, right?" Teller says in his opening. "But it's a little less usual to have an ex-Marine on your team to help you track that progress. Having one of your prototypes get away from you and head off across the countryside when it's the size of a bus? Even less normal. Saying 'thank God we have that ex-marine so that he can chase it down and take it down with a bowie knife—memorable! It sounds funny if you're not part of the Project Loon team at Google X, but it was an uncomfortable event for them. We now tie down our balloons as we're inflating them."

Teller, who has worked as a scientist and as an entrepreneur, holds a Ph.D. in artificial intelligence from Carnegie Mellon University and numerous U.S. and international patents related to his work in hardware and software technology. He was co-founding CEO of BodyMedia, Inc., a Pittsburgh-based developer of wearable body monitoring systems that was acquired last year by Jawbone, and co-founding CEO of Cerebellum Capital, Inc., a San Francisco-based investment management firm whose investments are said to be "continuously designed, executed, and improved by a software system based on techniques from statistical machine learning." Mincing no words, he tells the audience that "all things being equal, making hardware is just harder than making software-only kinds of things."

Each Project Loon balloon uses 2.4 ghz spectrum and can cover a 40 km diameter area below.
Photo courtesy of Google.

"You need a wider diversity of skills and you have to spend longer; you have to spend more money. And because the time from thinking of an idea to testing it and learning and repeating—that loop is the inner loop; it's the rate limiting step on innovation, and that's slower in hardware."

Why, then, would Google, which Teller describes as "one of the world's most iconically software-oriented companies," create Google X, a place dedicated to making physical products and services? Beyond the simplest answer—Google is already overflowing with brilliant and creative people that are tackling the software problems of the world—Teller offers a deeper reason.

"We say that we're in the business of making moonshots. By that, we mean finding—and, hopefully, over time, solving—really huge problems of the world, and doing that via radical technology solutions. Most of the world's really pressing problems, though, are physical in nature and require physical solutions," he explained. "So if anyone wanted to fix the clean water problem, and wanted to go after that using just software, I wouldn't give them great odds because it's just an inherently physical problem. If you want to take on grid scale energy storage so that renewable energy can really take off, it's such a physical thing, [and] if software is the only thing in your bag of tools, again, I'm not going to give great odds for success there. There are just a lot of these problems that are basically physical things or require physical systems in order to solve them. And the same thing is true for transportation, agriculture, for manufacturing, for telecom, and for infrastructure for healthcare."

Teller said that substantial cost and time pressures on businesses these days have led many of them to try to solve problems through software, which typically yields solutions that can be scaled up more quickly, less expensively without a lot of variable costs, and, ultimately, more lucratively. This tendency to use software as the solution is seen as a bias that leaves much fertile ground for innovation—what he called "green fields for exploration"—in attempting to solve the largely underexplored problems that aren't amenable to being solved solely by software.

"If you want to go solve a problem in the world that people have so far been unsuccessful at solving, planning to be smarter than them, planning to have more money than them, planning to work harder than them—these are not really great plans," he said. "The right plan is go look somewhere that they didn't even bother looking. So one of the reasons that Google X is focused on the physical world is because that bias tends to drag us into explorations where people have done less exploration in the past, and that, by itself, gives a sort of surprising boost in opportunities to find solutions—at least potential solutions—to really big, hard, interesting problems."

Most real breakthroughs in any problem-solving exercise, Teller said, involve breaking at least one assumption or constraint around a given problem. "By breaking those assumptions, those constraints, you have taken a problem that didn't look solvable and turned it into one that at least might be solvable," he explained. One might encounter a problem, for example, where half of the constraints and assumptions around the problem are physical in nature, and half of them are more algorithmic or data-oriented in nature—what he described as "the half-and-half problem." In such a case, a software-only bias can greatly inhibit the development of a solution. "If your team is determined from the beginning only to break the assumptions which are software-only, you have given up, from the beginning, half of the opportunities to win," he said. "If you've taken them off the table from the beginning because you don't have the skills, because you're not committed to the physical things, to doing those physical things, you just took down by a factor of two your chances of finding a solution."

Looking ahead, Teller foresees an evolution of physical systems over time that moves away from the kind of mechanical thinking that prevailed in the 20th century, toward an approach that's more open to new ways of thinking about the interaction between hardware and software. He sees it as a movement away from a foundational way of solving problems, "the civil engineering mentality," which has built itself into nearly everything in our lives—not just bridges, he said.

"Cars, for example, smart as they are, still fundamentally are built around this mechanical way of thinking. And I believe that in the end, what we're going to see is that more and more strength, flexibility, safety, and durability is going to be driven not by improving the mechanics of the system, but by embedding intelligence in the system, which means a physical system where the hardware and software are increasingly married in interesting and productive ways," Teller said.

A good example is Project Loon, an audacious attempt "to bring the technologies of access to everyone on the planet," said Teller in a YouTube video describing the project. Although the Internet is often termed "a global community," two-thirds of the world's population still lacks access to it. Project Loon seeks to address the problem via a network of balloons that travel around the globe, riding the high-altitude winds of the stratosphere. The network is "designed to connect people in rural and remote areas, help fill coverage gaps, and bring people back online after disasters," Google says on its website. (

Because the many layers of wind in the stratosphere vary in speed and direction, the balloons could very easily become separated from each other if not for the use of software algorithms that help keep them on track. Project Loon moves each balloon into the layer of wind that's flowing in the right direction, arranging the balloons to form a communications network that people can connect to via a special Internet antenna attached to their residence, according to the website.

"If you want a balloon that just sort of hangs over San Francisco and helps create Internet connectivity, you can't have it," said Teller. "You can't have that balloon stay there because the winds are too strong and you can't fight the winds; there's not enough energy up there to fight the winds. That would be a mechanical way to think about solving the problem. But if, instead, you allow the balloons—which is what Loon is doing—to go up and down, to pick winds of different directions and speeds, and by picking those different winds, to sail with the winds, then altogether, they can take advantage, by flocking behavior, of what no one of them could do. So you can think of a virtual balloon staying still over a particular place on the earth, even though no one balloon is staying there. That's embedded intelligence."

Another Google X project, the Makani airborne wind turbine (AWT), was on display at SolidCon and is one of Teller's favorite examples of software merging with hardware. Unlike conventional, stationary wind turbines built of heavy steel, the AWT is a lightweight, flying wing that is said to eliminate 90 percent of the material used in conventional wind systems. Similar to a kite, the wing is fixed to the ground by a flexible tether that reportedly allows it to fly in large circles at up to 1,000 feet altitude, where the wind is stronger and more consistent than the winds reached by conventional systems, according to the company's website. (

"This is explicitly a trade of 3 or 400 tons of steel being swapped out of the equation in exchange for an incredibly hard control systems problem," said Teller. "But what a wonderful tradeoff that is, because steel isn't getting lighter very quickly; it's not getting cheaper very quickly, but control systems are roughly following Moore's Law. Awesome trade!"

See the video of Astro Teller's keynote presentation at O'Reilly's Solid Conference 2014 here:

Home |  About Us |  Back To Technical Library |  Contact Us
Copyright © 1996-2010 All Rights Reserved.
General or Technical Questions? E-mail