ECE ILLINOIS Facebook ECE ILLINOIS on Twitter ECE ILLINOIS Alumni and Friends on LinkedIn ECE ILLINOIS Instagram

Contact Info

Brad Petersen
Director of
Communications
1066 ECE Building
306 N. Wright Street
Urbana, IL 61801
Phone: (217) 244-6376
bradp@illinois.edu

Contact Info

Meg Dickinson
Communications Specialist
1068 ECE Building
306 N. Wright Street
Urbana, IL 61801
Phone: (217) 300-6664
megd@illinois.edu

World's fastest quantum random-number generator

 Subscribe to ECE ILLINOIS News

By Celia Elliott, Department of Physics
June 2, 2010

  • New method of generating random numbers at record speeds has been developed by Illinois researchers.
  • ECE graduate student Michael Wayne worked with Physics Professor Paul Kwiat on the problem.
  • The method shows promise for use in encryption, statistical analysis, and advanced numerical simulations.

Ever since humans discovered gambling, people have sought improved means of generating random numbers—unpredictable outcomes based on a physical process such as coin flipping, dice throwing, or wheel spinning. But such methods are both too slow and too unreliable for modern applications requiring random numbers.

ECE graduate student Michael Wayne (left) and Physics Professor Paul G. Kwiat at work in Kwiat's lab. Photo by R. Hohenstein.
ECE graduate student Michael Wayne (left) and Physics Professor Paul G. Kwiat at work in Kwiat's lab. Photo by R. Hohenstein.

Now researchers at the University of Illinois at Urbana-Champaign have developed a novel method to generate random numbers at record speeds and security, using the laws of quantum mechanics. The new scheme, based on shaping the photon flux from a laser diode and then digitizing the time interval between random photon arrivals, is a factor of 10 faster than any other quantum random number generator reported so far, according to Bardeen Professor of Physics and of Electrical and Computer Engineering Paul G. Kwiat, an ECE affiliate. The group’s results were published in Optics Express in April.

Random number generators are essential for a variety of applications, including data encryption, statistical analysis, and advanced numerical simulations. However, because of limitations in reading out truly random physical processes, many current applications employ a pseudo-random number generator—a deterministic method that replicates the behavior of a physical phenomenon that is expected to be random or a computational algorithm based on a shorter initial value, known as a “seed” or a “key.”

But some applications, such as quantum cryptography, require absolute randomness to ensure security. Explains ECE graduate student Michael Wayne, who developed the new method as part of his graduate research, “Most random number generators are not actually random, they are just so complex that the computational cost required to predict their outcome is too large for modern computers. As technology advances, this is no longer the case, and previously secure systems can be compromised.” Because quantum physics is intrinsically random, scientists have increasingly turned to quantum systems as a source of random data.

Quantum optics, the behavior of individual “particles” of light, called photons, has proven to be particularly amenable to generating and reading out the random binary numbers of great interest for secure information processing, encryption, and transmission. Most existing quantum random number generators rely on measuring the behavior of an incoming photon at a beam-splitter to create data. This approach has significant limitations, however, in that each photon can produce at most one bit of data, and the systems are heavily constrained by the rate at which single-photon detectors can operate.

The method developed by Kwiat’s group produces a fast quantum random number generator having reduced bias and requiring less post-processing. “Unlike existing methods, our method creates multiple random bits per detection event and greatly reduces the need for post-processing,” said Kwiat. “We are able to obtain fast, secure quantum random number generation at rates exceeding 100 Mbit/s. Even faster rates—exceeding 10 Gbit/s—may be possible with planned improvements to our laser driver circuit and detectors.”

Editor's note: media inquiries should be directed to Brad Petersen, Director of Communications, at bradp@illinois.edu or (217) 244-6376.

 Subscribe to ECE ILLINOIS News