Read the Original Article at http://www.informationweek.com/news/showArticle.jhtml?articleID=240142639
Here's a stat sure to become PowerPoint porn in the months ahead: General Electric predicts that the "industrial Internet" could add $10 trillion to $15 trillion to the world economy in the next 20 years.
Indeed, $15 trillion is a "wow, that's big" number sure to be dropped into many a presentation, but it's not the most important part of GE's major new report on its industrial Internet vision. The most important part is why GE would bother to calculate this projection and issue such a report. The reason -- beyond the marketing value -- is that GE needs a whole lot of help from other vendors, regulators, financiers and users of technology before this vision and its $15 trillion payoff can come true. This report looks like an attempt to rally an ecosystem.
GE describes an industrial Internet where the machines it makes, such as jet engines, power plant turbines and MRIs, constantly gather data and send it along over the Internet for analysis. The data might alert people to take action, like replace a part that's close to wearing out, or tell a machine to automatically take an action, like slow down a turbine that isn't needed. The idea's more broadly discussed as the "Internet of things."
GE created the report in part because the whole idea of Internet innovation is under attack, says co-author and GE chief economist Marco Annunziata, and "we thought it was important to challenge this view." The report cites Northwestern University professor Robert Gordon for his view that the "innovations of the Internet Revolution are simply not as transformative as those of the Industrial Revolution." People think of the Internet's impact as centered on entertainment and social networks, Annunziata says. That's why GE spends much of the report trying to quantify the industrial Internet's potential in terms of economic and productivity growth.
But this industrial Internet needs a lot of technical and regulatory pieces to expand at the scale GE envisions. Based on my reading of the report, my interview with Annunziata, and InformationWeek's reporting on the Internet of things throughout this year, I think the following changes must happen for GE's growth vision to become reality.
This is No. 1 on Annunziata's list of required technology advancements. The cost of sensors is dropping, and the reach of wireless and wired Internet is expanding. It's getting cheaper and easier to gather and transmit data. Data analytics capabilities must improve "to make sure the data that comes available can be used in a productive way," Annunziata says. In GE's report, he describes this capability as "harnessing the power of physics-based analytics, predictive algorithms, automation and deep domain expertise" to know how machines and systems operate.
Startups and established vendors are developing big data analytics to make sense of the Internet of things, and do so quickly enough to make a timely business decision. GE's own software unit is among them. In an article earlier this year, here's how we described one technology problem GE software is working on, that of combining real-time data and what-if scenarios:
What GE wants to offer is the ability to ask, "Has any machine in our entire system ever had X, Y and Z factors, and what happened four hours later?" GE's systems today would take about 30 days to answer that question -- if they could even answer it. GE's working to combine its data management and analytics software with Hadoop-based data processing to deliver an answer in 30 seconds.
Automation Must increase
As companies collect more data from more sources and then try to make faster decisions with it, the complexity soon exceeds humans' ability to keep up, Annunziata says. People must continue to oversee the process, he says, but more machines need to automatically take actions. Automation requires more embedded technology and, back to analytics, better decision-making software. Automation also points to the dark side of the industrial Internet: Doing it right means destroying a lot of manual jobs and betting that economic growth creates enough new ones.
People Need New Skills
Annunziata highlights three new job growth areas: skills that cut across traditional lines of engineering and software development, creating a new role like a "digital-mechanical engineer"; data scientists, who specialize in fields from cybersecurity to pattern recognition to data visualization; and user interface experts, who can design human-machine interfaces that make a job easier and people more productive. We recently wrote about how Ford is recruiting more electrical and software engineers to do this kind of engineering-plus-design work for its cars, as software becomes a bigger part of why people buy a vehicle. As cars get more connected, such as sharing data car-to-car to know if there's an accident or traffic jam ahead, these skills get more important. Companies such as GE and Ford need universities to start training such specialists.
Policymakers And The Public Must Be Convinced
How much machine automation should people allow? We see this debate beginning around Google's self-driving car. In financial markets, automated high-speed trading creates controversy at times such as the "flash crash," when markets seem to overreact.
The risks are real, Annunziata says, so we need "transparent public debate on how much control we are giving to machines." Cybersecurity also becomes of vital public importance when it relates to networked power plants, jet engines and healthcare equipment. GE argues for cybersecurity regulation that's less fragmented across states and countries. Again, it points to why GE needs to make a case that the risks and economic disruption of a more networked and automated business world will pay off in economic growth.
Companies Must Invest
GE's report focuses mostly on macroeconomics. Getting the U.S. economy back to the 3.1% productivity growth rate of the 1995 to 2005 Internet boom, rather than the 1.6% rate since then, drives its prediction of $10 trillion to $15 trillion in economic growth from the industrial Internet.
But investing in the industrial Internet is a microeconomic issue -- companies make this decision one by one, project by project. The New York Times, in writing about GE's industrial Internet concept, cited an example of a wind farm operator upgrading the sensors and optimization software -- and netting a modest 3% energy output gain. InformationWeek wrote about Union Pacific's system to monitor train wheels and use analytics to predict failures, leading to a 75% drop in wheel-related derailments, but the next level of investment and innovation hinges on more effective sensors, better predictive analytics and better data sharing to predict the effects on the entire rail network.
The decision by companies whether to invest in this technology brings us back to the need for better analytics, more automation and new skills, factors that will drive the ROI of these Internet of things projects. We are seeing companies take these steps and make incremental gains with networked machines. But we need an innovation ecosystem to kick in to get anywhere near GE's $15 trillion vision.