By JULIE MORAN ALTERIO
jalterio@thejournalnews.com
THE JOURNAL NEWS
What do the hard disk drive, a supercomputer smart enough to beat a world chess champ and the excimer laser surgery technique that's fixed myopia in millions have in common? All were invented in the research labs of IBM Corp.
Tuesday marks the 60th anniversary of the opening of the Watson Scientific Computing Laboratory. What started in a renovated fraternity house near Columbia University in New York has become the largest corporate research division in the world, with 3,000 scientists toiling at eight labs in six countries.
Over the years, the original mission to build calculating machines to aid the war effort has morphed into a wide-ranging endeavor to invent the future of computing, from the mighty Blue Gene supercomputer that's designed to simulate the complex action of proteins in the human body to minuscule carbon nanotubes that could one day replace silicon in computer chips.
Physicist Jia Chen joined IBM Research in 2003 and today is studying carbon nanotubes at the division's headquarters in Yorktown.
Carbon nanotubes — 50,000 times thinner than a human hair — are one of the most promising materials scientists are pursuing to build smaller and smaller computer chips.
Even so, the 33-year-old Ossining resident said it might take 10 to 15 years for carbon nanotubes to make it into a microprocessor — a timeframe she's happy with given the import of what she's researching.
"We could have a significant contribution to a future technology that could impact everybody's life," Chen said.
Though 50 years separate Chen from Fran Allen in age, the elder researcher voices the same passion about her work at IBM.
Allen joined IBM in 1957 to work at a laboratory in Poughkeepsie, where her colleagues were just inventing the field of computer programming.
"We made a lot of mistakes, and we invented a great many things," she said.
Allen's training as a math teacher wouldn't earn her a slot as a programmer at IBM Research today, but back then there weren't computer science departments at colleges.
Allen was hired with three other women, and the four rented a house nearby in Wappingers Falls that became both a social center and a sanctuary where colleagues gathered to discuss the computer systems they were creating.
"So many of us were single and young, there were no boundaries between work and home. It was a tremendously productive period for all of us," Allen said.
Ski trips, hiking trips and just hanging out were an excuse to work out problems. "Napkins were a favorite tool for doing design," she said.
Allen's work on optimizing computer code is considered pioneering in the field. The Croton-on-Hudson resident was the first woman to become an IBM Fellow, the company's highest technical honor.
"It was very, very challenging. There was never a doubt that it was important," said Allen, who retired in 2002.
Working at a corporate research lab such as IBM's was satisfying, Allen said, because she could see the results in actual products.
"The role of research has always been to do the groundbreaking work and, as part of IBM, to transfer the work to products," Allen said. "I see my career as standing on a wall and seeing both ways. One way is seeing new possibilities and the other is building new products."
Bill Strachan, the program director for technical recruiting for IBM Research, said that mix of pure science and product development makes a career at the company appealing for Ph.D. graduates.
IBM hires about 100 new researchers a year.
"It's intriguing to students that researchers will meet customers 10 to 20 percent of the time. They are looking to solve real problems," he said.
That interaction has intensified in the past 10 or so years as IBM has increasingly viewed its research division as a competitive advantage. A staff of scientists who aren't contributing to the bottom line is a luxury the company can no longer afford.
"Typically we were the ivory tower, but we are an arm of the IBM Corp. and we need to benefit the IBM Corp.," Strachan said.
This strategy is also designed to avoid a repeat of missteps of the past. IBM's invention of relational databases and Reduced Instruction Set Computer architecture benefited competitors such as Oracle and Sun Microsystems years before IBM moved into the database and Unix computing markets.
The head of IBM Research, Paul Horn, senior vice president, said there is a sense in IBM that his division is what sets the Armonk-based computer giant apart from companies that compete on price.
"IBM's business model is to operate in the high value segment of the information technology market. High value means high margins, and it means you can get a return on the research and development dollars you spend," Horn said.
Since 1996, IBM has invested about $5 billion a year on research. Last year, the company spent $5.7 billion.
That investment has translated into a boom in patents. IBM has been granted more U.S. patents than any other company for the past 12 years. Last year, IBM received 3,248 patents.
Increasingly, more of the ideas being generated are related to IBM's software and services business, said Horn, who joined the company 26 years ago.
"When I came here, we were 85 percent or more hardware. There were very few IBM fellows in software. Software was just stuff to make the hardware work," Horn said.
Today, researchers are turning their attention to consulting with clients to solve complex business problems such as managing fleets of vehicles or analyzing text on the Web.
This shift is a reflection of the contribution of computer services to IBM's finances, with $46.2 billion of $96.3 billion in sales last year generated by IBM Global Services.
"We've staked out as our business model to operate in those areas of the IT industry where research can make a difference, which is pretty nice for the research team," Horn said.
Frederick P. Brooks Jr., a professor of computer science at the University of North Carolina, said the challenge for IBM and other corporate research labs is twofold: "How to fund and maintain a basic research program in hard economic times, and how to get great ideas to transfer effectively to the product divisions," he said.
Brooks will be in Yorktown on Tuesday when IBM celebrates its anniversary to lead a discussion on the research division's early days.
Brooks contends that IBM could not have become the company it is without its research division.
"IBM has maintained technical leadership in many areas — magnetics, silicon processing, circuits, computer architecture, compilers — in large part because the research division was well out in front of the developing technologies, mastering them before it was economically or technically feasible to put them into products," he said.
IBM's creation of a research lab reinforced the example of such rivals as Bell Labs and stimulated others such as Xerox's PARC and RCA's Sarnoff Center, Brooks said.
Though the Watson lab wasn't started until 1945, IBM had begun experimenting with a research center more than a decade earlier.
In 1933, during the Great Depression, Thomas J. Watson Sr. oversaw the creation of a dedicated center for the company's engineers and inventors in a building near a manufacturing plant in Endicott, N.Y.
Engineers there researched new types of insulated wire for electrical connections and studied how to strengthen paper for the company's punch cards.
Journalist Kevin Maney, author of "The Maverick and His Machine," a biography of Watson Sr., wrote that IBM spent $1 million — or nearly 6 percent of annual revenue at the time — on that first lab.
The lab was stocked with the latest scientific instruments and a basement weather room that simulated temperature and humidity in every climate where IBM machines were used.
"How could Watson justify this? He spent $1 million for something as amorphous as research and development when businesses were falling apart and nearly one-quarter of Americans had barely enough to subsist. But building the lab had a logic to it. Somehow, Watson had to stimulate demand. He had to come up with products that companies couldn't resist, whatever the economic conditions," Maney wrote.
The breakthroughs of that lab included the first punch card machines that could handle multiplication and division.
Today, the pre-eminent machine at IBM Research is the Blue Gene supercomputer, the product of a $100 million, five-year effort begun in 1999.
A Blue Gene system built for the Department of Energy's Lawrence Livermore National Laboratory in California is the fastest computer in the world, capable of 136.8 trillion calculations per second, or 136.8 teraflops.
The second-fastest — at 91.2 teraflops — is a Blue Gene system installed in Yorktown.
Al Gara, chief architect for Blue Gene, said designing a new machine with the ability to tackle the complexity of human biology meant rethinking the way supercomputers are built.
"We were trying to do some groundbreaking in terms of protein folding but also in terms of the computer science," Gara said.
Rather than build a power-hungry custom system, IBM scientists used a low-power computer chip in a new configuration. The result? The Blue Gene system for Livermore consumes 15 times less power and is at least 50 percent smaller than the fastest computers of just a couple years ago.
"It's quite clear that the way to get more performance in the future is unrelated to how we got more performance in the past," Gara said.
As the limits of miniaturization of silicon transistors approach, research into new methods to build faster computers is vital, Gara said.
"The future is really unknown, and there is enormous activity and excitement at the lab. If ever we need a research division, it's now," Gara said.
Ensuring there will be an ample supply of scientists to staff the lab is also a priority as IBM ponders its future, said Jim Wynne, a Yorktown researcher and manager of the company's outreach to schoolchildren.
Wynne — who was inducted into the National Inventors Hall of Fame thanks in part to his role in developing excimer laser surgery — said about 300 of his colleagues lecture students on everything from nanotechnology to how search engines work in an effort to spark their interest in science.
"I'm not going to be able to invent the rest of my life. I want to see a future generation of youngsters from my community continue in this field. IBM wants to invest in the pipeline."
IBM Research highlights 1945 1956 1957 1961 1966 1967 1970 1980 1981 1986 1992 1997 IBM opens Watson Scientific Computing Laboratory in New York City.
IBM researchers in San Jose, Calif., produce the first magnetic hard disk. These disks are a vital part of all computers today. Researcher John Backus invents FORTRAN, which became a widely used computer programming language for technical work. The language allowed programs to be written in an easy-to-understand format, such as C=A/B, instead of in 1s and 0s.
IBM moves its research headquarters from Manhattan to Westchester County.Yorktown inventor Bob Dennard invents one-transistor dynamic RAM, known as DRAM. Chips based on this invention are still the dominant form of computer memory.Yorktown researcher Benoit B. Mandelbrot publishes a paper introducing fractal geometry, which describes the shape of irregular natural objects, such as tree branches.
IBM scientist Ted Codd publishes a paper introducing the concept of relational databases, which stores data in tables that are easy to interpret by nontechnical users. Nearly all databases today are based on the concept.
IBM builds its first prototype computer using Reduced Instruction Set Computer architecture, which was invented by IBM scientist John Cocke in the 1970s. Two scientists from the IBM Zurich research laboratory, Gerd K. Binnig and Heinrich Rohrer, invent the Scanning Tunneling Microscope, which provides a first-ever look at the topography of atoms. The scientists won the Nobel Prize in physics in 1986 for this work.
Two scientists from the IBM Zurich research laboratory, J. Georg Bednorz and K. Alex Muller, discover superconductivity in ceramic oxides that carry electricity without loss of energy at much higher temperatures than any other superconductor. A year later, the men win the Nobel Prize in physics for their discovery.
IBM introduces ThinkPad laptops, which feature the TrackPoint, a little red pointing device in the middle of the keyboard that was created by IBM researcher Ted Selker, who is now at MIT.
Deep Blue defeats World Chess Champion Garry Kasparov in the first known case of a computer besting a world class player in a tournament.
No comments:
Post a Comment