Prucnal making light work to accelerate the Internet
Princeton NJ -- It's been 20 years since Paul Prucnal saw the light.
It was before the Web, around the time of the first PC and when the Internet was just a specialized tool for universities and government labs. No one was downloading videos or clogging networks with e-mail, but Prucnal, a professor of electrical engineering, was looking for ways to make computer communications faster. And the answer, he knew, was light.
But Prucnal, then at Columbia University, wanted to push the idea much further. The real bottlenecks, he realized, were not in the long hauls. The thin strands of glass webbing the nation were like giant water mains spilling their contents into inadequate local plumbing. The information in the light pulses had to be converted back into cumbersome electrical signals before they could be sorted and directed to their proper destinations.
Prucnal's idea was to build "all-optical networks" in which all the electronic sorting and routing equipment would be replaced by devices that deal only in light, with no conversion to electrical signals.
At the time, said Prucnal, it was a "pie-in-the-sky, crazy idea." No such device, or even the materials necessary to build it, existed. It was a little like trying to make a hand-held FM radio before the invention of the transistor.
"People didn't laugh at it," Prucnal recalled, "but they recognized that it was way in the future."
Today, Prucnal has solved several of the key problems. He invented an all-optical, microscopically small switch that is undergoing commercial development and appears to be on the verge of widespread use. With the Internet having undergone radical growth and the demand for high-speed networks soaring, his 20-year-old idea may be just in time.
From billions to trillions
Sending information through a fiber optic cable is a little like sending Morse code with a spotlight, except that the code is binary computer language and the flashes are only a billionth of a second long. By adjusting the timing of the pulses and the wavelength of the light, engineers have learned to transmit more than a trillion bits of information per second through a single fiber. That's the equivalent of 200 CDs full of music each second.
When all that information hits a junction in the Internet, it needs to be sorted and forwarded along the proper channels according to a tiny amount of address information tacked onto the front of each packet of data. The process is called routing. The fastest electronic routers in use today can handle only about a billion bits per second. That means incoming data is often stored until the electronics catch up. Prucnal believes that a router based on his all-optical switch could handle 1,000 times more data and keep up with the trillion bits per second coming in from fibers.
This optical router design has sparked interest among major equipment manufacturers such as Cisco Systems and Sun Microsystems, Prucnal said, noting that such optical systems may come into regular use within the next five years.
Prucnal also is part of a multi-institutional project funded by the U.S. Department of Defense to develop the technology for the next generation Internet. The project, called Pegasus, aims to make the Internet 1,000 times faster while supporting 100 times the current number of users and applications. For that project, Prucnal is putting the same switch to use as the core component of two devices: an optical regenerator, which boosts the intensity of the light pulses and restores their shape; and an optical converter, which changes light from one frequency, or color, to another.
Prucnal's switch, called TOAD as a benign acronym for the technical terms "Terahertz Optical Asymmetrical Demultiplexer," has many other potential uses beyond telecommunications. In one project, Prucnal and his students are developing the switch for use in a medical imaging device that could use light to probe tissue, similar to the way in which a sonogram uses sound waves but with much higher resolution. A laser would shine light through the skin and the TOAD device, with its ability to respond to very rapid pulses of light, would record the scattered reflections that come back.
"One optical pulse can give you a tremendous amount of information about what is inside of the tissue," Prucnal said. One use would be to look for very small tumors that are relatively near the surface as in breast cancer.
He also is working with engineering professor James Sturm and the University's Center for Photonic and Optoelectronic Devices to develop a related technique that would use pulses of light to analyze individual DNA molecules.
Each of these applications remains five or more years in the future, but the combination of basic research and real-world applications in Prucnal's work is very appealing to students, and has helped make him one of the most popular teachers in the department, according to department chair Peter Ramadge.
Prucnal recently took over the department's introductory class on circuits, which is required for all electrical engineering majors, and also teaches a senior-level course on optical communications. The latter draws about 40 students a year, which is very high for such a specialized senior-level course, Ramadge said.
"He's very well liked by the students," said Ramadge. "He takes a keen interest in them, and he has a style that encourages them to get involved and do projects and try things for themselves."
"We weren't just doing esoteric things," said Robbie Saperstein, who graduated last year after working on two independent projects with Prucnal. "You could see the applications and where it was going."
Saperstein, now a graduate student at the University of California-San Diego, credits Prucnal with making his transition to graduate school easy. The cutting-edge lab work made him comfortable moving straight into independent work, and Prucnal's personal guidance played an important role in his choice of schools. "I owe a lot to him," Saperstein said.
Bob Schell, who also graduated in 2001 and is now working for a fiber optic networking company in New Jersey, also said that work in Prucnal's lab prepared him well for addressing difficult problems even if they are not directly related to ultrafast optical switching.
"He's extremely personable," said Schell. "He's always there. If you want to just drop by, it's easy to find him. He takes a lot of time (with us)."
Prucnal tries to keep students in his lab focused on projects that are about 10 years ahead of everyday use, but not so advanced that they have no connection to the real world. As research comes closer to real-world use, he looks further to the future for interesting problems that are still in the category of "pie-in-the-sky."
Prucnal and his students, for example, are now trying to understand if there is a fundamental physical limit to how fast optical switching can be done and where that boundary might be -- not that they are inclined to accept limits. At switching rates faster than a trillion times a second, very small-scale effects cause the materials to behave chaotically. Prucnal, however, is trying to understand these effects and is looking for ways to exploit them to push switching speeds even faster.
"It's fundamental. It's exciting. There's absolutely no
need for it in the communications world these days," he
said. "But we are definitely looking toward the future."
Deadline. In general, the copy deadline for each issue is the Friday 10 days in advance of the Monday cover date. The deadline for the Bulletin that covers Dec. 3&endash;9 is Wednesday, Nov. 21. A complete publication schedule is available at deadlines or by calling (609) 258-3601.