Friday, December 28, 2012

TECH SPECIAL….DREAM MACHINES : GE’S “INDUSTRIAL INTERNET”


DREAM MACHINES :  GE’S “INDUSTRIAL INTERNET”

How GE ‘s over $ 100 billion  investment in    “industrial internet”   will add $ 15 trillion  to world GDP

http://articles.economictimes.indiatimes.com/images/pixel.gif
http://articles.economictimes.indiatimes.com/images/pixel.gif
Sundar Murugappan is a quintessential Chennaiite who talks briskly as if in no mood to stop unless interrupted. The 27-year-old is a PhD from Purdue University where he spent many months studying "interactions" between human beings and computers, an area of research he says he loved the most.
He joined General Electric Co (GE) less than a year ago at its Global Software Center in San Ramon, California, as a "user experience researcher", a job that entails him to collect a lot of data besides writing software and doing surveys.http://articles.economictimes.indiatimes.com/images/pixel.gif
In the past several months, Murugappan and his colleagues have collected data on power outages in hospitals in parts of the US to help predict and avert outages, including those caused by "vegetation" on the streets. Currently, he is working on a health-care project where collecting data from hospitals means helping patients cut costs.
His suggestions include redesigning "patient experience" to avoid unnecessary treatment because a "good chunk" of those who get admitted to hospitals are those who just happen to be "hyper" when they arrive at hospitals for consultation.
"My work is simply exciting and new. I am where you do the kind of academic research that impacts the lives of people directly," says this BITS Pilani alumnus.
By collecting data, GE expects to achieve many more things, and some of them are averting fires, machine breakdowns and air crashes; forecasting natural disasters with much greater accuracy; and so on.
Internet Gets Physical

What's a dream job for Murugappan and others is GE's big bet, not only to sustain its top-of-the-league place in world business, but also to redefine the industrial process.
That's why GE, a $150-billion MNC, has planned an investment of more than $100 billion and hired more than 400 engineers at its San Ramon centre  where the project is to develop digital tools to analyse millions of gigabytes of data generated by machines.
In a report it released late last month, the company expects the process it has kick-started connecting machines to the internet  to generate an additional $15 trillion in global GDP by 2030 by helping to trim costs and wastages. GE calls it the "industrial internet", to drive home the difference with consumer internet where people, not machines, are connected to the World Wide Web.
In fact, we are already seeing the impact of industrial internet, says Chris Murphy, editor, InformationWeek: "This is real technology that companies are using today, not science fiction."
He goes on: "We're calling it more broadly the internet of things... look at John Deere Tractors that lets a dealer remotely diagnose problems using sensors on the vehicle tied to a wireless link. Union Pacific railroad has sensors on its tracks that take 20 million readings a day to predict when a wheel is at risk of failing. Vail Ski Resort has RFID [radio frequency identification] tickets that track skiers and let them know how many vertical feet they covered in a day," he explains.
Doors of Perception
GE sees its efforts to create tools for analysing big data (see The Concept) to help slash costs and enhance efficiency across sectors (see Estimate of Gains) along the lines of how undersea cable revolutionised communication and steam engines transformed travel. Also, the company expects the software business, which posted revenues of $2.5 billion last fiscal, to see double-digit growth until 2015.
Its advantage is that analysing big data from machines is not everyone's cup of tea, says Silicon Valley veteran William Ruh, who GE hired from Cisco last year to set up the San Ramon project."GE is one of the companies globally that have the ability to do that," he says.
GE makes machines of many types. Besides, from five billion devices that are connected to the internet now, 2020 is expected to see as many as 50 billion devices — many of them industrial machines — connected to the Net, according to reports.
Tim O'Reilly, founder of O'Reilly Media, says Ruh's pride isn't misplaced. "[Thanks to the industrial internet], there is a growing sense in the global tech community that we should be doing much more important work than filtering social gossip to target ads." He adds: "Our world has higher priorities and we want to be involved in solving them.
It crystallised when [former Facebook veteran and data expert] Jeff Hammerbacher rued, 'The best minds of my generation are thinking about how to make people click ads'." O'Reilly notes that the latest issue of MIT Technology Review has captured that feeling: the magazine has a picture of Buzz Aldrin on the cover with the line, "You Promised Me Mars Colonies. Instead, I Got Facebook".

The November-end report by GE, authored by Peter C Evans, its director of global strategy and analytics, and Marco Annunziata, its chief economist, says that the full potential of the industrial internet will be felt when the three primary digital elements — intelligent devices, intelligent systems and intelligent automation — fully merge with physical machines, facilities and networks.
Too good to be true?
DJ Patil, data scientist at Silicon Valley-based venture capital firm Greylock Partners, shares Murphy's views that a beginning has already been made in examining big data, not just from social networking sites but also from machines. "People are already using big data to make better decisions and improve processes... yet I'm excited about the industrial internet. The new white paper from GE outlines the expected benefits quite well.http://articles.economictimes.indiatimes.com/images/pixel.gif
For example, 1% increase in fuel efficiency for airlines saves $30 billion over the next 30 years," contends Patil who, along with Hammerbacher, coined the word data scientist, which, according to Harvard Business Review, is "the sexiest job of the 21st century". A data scientist, Patil says, is someone who "finds rich data sources, works with large volumes of data, cleans the data and makes sure that data is consistent, melds data sets together, visualises that data and builds tools that enable others to work with data effectively".
Ruh says that's exactly what his team — including freshers such as Murugappan — is doing at GE, with the aim of solving real-world problems. O'Reilly points out that such problems include energy sustainability, transportation congestion and so on. "These are hard problems with huge payoffs and that makes them interesting," he says.
Industrial Revolution 2.0
GE, founded by iconic scientist Thomas Alva Edison, has the habit of being right about biggish trends for over a century, from electricity to radio to computer and even back offices.
According to this sixth-largest US company by revenues , it is the only company that was on the original Dow Jones Industrial Index and is still there; the next best thing is the industrial internet. "It will be a game-changer," says Ruh, echoing chairman Jeffrey Immelt's vision for the conglomerate.
GE expects the industrial internet to change the world as much as the Industrial Revolution of the 18th and 19th centuries did. Well-known American mathematician Brian Conrey finds it heartening.
He forecasts that "the higher-level, more focused interactions" of industrial internet will enable businesses of all sizes to compete on a level-playing field just as the internet itself empowered individuals. For O'Reilly, this is an opportunity to bring "what Jonathan Zittrain refers to as the internet's 'generativity' to the industries that make things ; different kinds of solutions to industrial problems".
Conrey is of the view that the "ongoing revolution" will have a big effect on the world, probably bigger than the Industrial Revolution itself. "Much as the Industrial Revolution was largely responsible for the birth of the middle class, we have already had glimpses of the power of the internet as a great equaliser," he says.
O'Reilly thinks a bit differently: industrial internet cannot be fully compared with the Industrial Revolution, he says. "The Industrial Revolution created the first sustained growth in human wealth in history, and that first can't be repeated. But there is still a lot of yes in my answer... as our things and machines follow us onto the network, and there are more of them than there are us, the opportunity for ad hoc coordination, optimisation and resilience is huge. During the next 20 years, as our things and industrial machinery colonise the Net, that's going to translate into tremendous productivity gains and economic performance," he hopes.
In fact, the internet revolution of the late 20th century raised American labour productivity growth to an average annual rate of 3.1% from 1995 to 2004, twice the pace of the previous quarter-century.
San Francisco-based software engineer Byju Sukumaran, who Oracle hired from Cisco last year to work on its business intelligence and big data projects, expects GE's big data project to be a harbinger of more ventures and investments in this space. Like others, he too notes that the concept is already around. "Smarter Planet” was IBM's industrial network.
Google, Yahoo! and Microsoft and the open source community had already begun taming big data," he says emphasising that GE's plan is still innovative enough. "It is in the right place to do what it is doing now since it produces and services pretty much everything from jet engines to home appliances to power plants and medical equipment," he says, adding that it will be very interesting to watch what comes out of it.
"I am optimistic that this will help them find ways to design better products, reduce costs and improve efficiencies," says Sukumaran, a Carnegie Mellon University alumnus who holds several US software patents.
Kenneth Cukier, the data editor of the The Economist and co-author of Big Data: A Revolution that Will Transform How We Work, Live and Think, dwells at length on why GE's entry into the internet of things (IoT) space is destined to make a huge impact; IoT has been defined as the idea that any physical object can connect to the internet and communicate with other objects to relay information to people.
The key takeaway, says Cukier, is that GE will stand to gain a lot as the industrial internet makes low labour costs less relevant for business advantage. What will become more important will be the capability to scrutinise data, he notes.
"Emerging markets have an advantage in '20th century things' like labour costs. But big data lets the West claim an advantage in the '21st century way' — one can become more efficient and productive by harnessing the data," he declares.
Adds he: "And this is why GE's entry [into the big data space] is so important. They can use their scale advantage that is today based on physical resources to a scale advantage for data." Cukier's soon-to-be-released book on big data he co-authored with Viktor Mayer-Schonberger explores how society is changing because of greater capabilities to analyse data.
Future Shock
Murphy hopes that industrial internet will have a decentralising effect on businesses. "What we are seeing today are micro uses [of IoT], where one company controls the whole information loop. There's a lot of work to do to make the networked systems people are envisioning." He also says that he prefers to think about the long-term idea that IoT would help relieve congestion in the world's megacities.
"Could self-driving cars that communicate with each other, with traffic lights and with the roads themselves help traffic move more efficiently, and ease the threat of 'global gridlock'?" wonders Murphy.
Like Murphy, Cukier and Conrey, O'Reilly too sees a level-playing field emerging, thanks to industrial internet. "With such large efficiencies at stake, it won't be just a company here and there, but entire industries that will move over to software control," he argues.
Such transformations are soon going to be treated as natural, says Sukumaran. His logic is that industrial internet is more than mere futuristic self-driving cars, blood pressure monitors predicting a heart attack, cell phones doing the shopping or devices forecasting fire in the neighbourhood, etc.
"It will end up massively impacting our economies," he says.
IoT Spell on India, China
Fall in costs and rise in efficiency may pep up economies, but the flip side is that there will be a shortage of talent necessary for organisations to take advantage of the surge created by industrial internet. According to McKinsey data, by 2018, the US alone could face a shortage of 140,000 - 200,000 people with deep analytical skills as well as 1.5 million managers and analysts with the know-how to use the analysis of big data to make effective decisions. Experts see talent crunch in emerging markets to be far worse.
They also see industrial internet, which will replace conventional ways, to cause job losses — something that may come in the way of its promises to generate extra trillions towards global GDP.
However, market watchers such as Ben Cavender, an associate principal with Shanghai-based market intelligence firm China Market Research Group, are far more optimistic. "It will likely lead to jobs transfer," he insists. He expects to see lower skilled technical and repair jobs being replaced by jobs that require greater facility with the Internet or diagnostic tools. "But overall job numbers are not likely to be reduced," he says.
Which means, says Cavender, industrial internet should benefit both China and India because both countries have large pools of young, college-educated workers to draw on to fill new jobs and both would ideally like to create more jobs for college graduates.
He also sees the two countries becoming leaders in innovation of new technology concepts over the next decade, with companies like GE increasingly targeting them for talent and to sell products.
"Because their economies are developing quickly there is also an opportunity to put in place the concept of industrial internet with less resistance than might be found in a more mature economy that has more systems in place," he says. John Flannery, president and CEO of GE India, also says that the net effect of the industrial internet will be positive. "As have virtually all productivity enhancers been in the past," he says.
There is still fear among people about intrusion into privacy and the overwhelming advantage GE would have in the new scheme of things. Conrey and O'Reilly warn against such generalisations. They say that as data sciences and maths become far more important for the world than they are now, smaller players can't be far behind.
Cukier explains why we are on the cusp: big data is going to give a quantifiable dimension "to things that are today measured by gut instinct or based on people's experience".
Take the example of doctors making diagnoses," he illustrates. "Over the course of a 40-year career, they may see 25,000 different patients. Over time, they gather more knowledge so that they're wise by the end of their career —they've seen lots of cases, and a few rare cases, in which the patient died but they now have a better sense of how to treat the disease. Now consider what big data means to their work.
By aggregating every electronic health record of every Indian citizen over the course of a decade, one can 'spot' what treatments work and what do not, which may surprise us. We can identify rare cases early on and know what drugs work best —based on far more 'rare cases' than any single doctor could ever see in a lifetime, or a hundred lifetimes, since it's based on many millions of cases."
Saving lives, creating jobs and adding trillions of dollars to economies — the future is here. And GE wants to get to that future first.
Existing Big Data Management Systems
Apache Hadoop is an open source software that enables the processing of large data. It is designed to scale up from a single server to thousands of machines, with a very high degree of fault tolerance. Others include HSQL, Cassandra, MongoDB.
Ullekh NP,  http://articles.economictimes.indiatimes.com/2012-12-16/news/35837159_1_ge-internet-data/

No comments: