Researchers cast wary eye on atomic-level computing.

Government researchers have spent a decade investigating computers that use the principles of quantum mechanics to process information much faster than currently possible. The promise of quantum computers is enormous, but experts within the U.S. military research community disagree on the scope and utility of this technology.

Experts point out that quantum computers could execute calculations several millions of times faster than conventional systems, but that the technology still is years away from becoming truly functional.

Interest in quantum computing is growing. Since 1994, the government's annual funding of quantum computing programs has gone from near-zero to $80 million a year, noted Henry Everitt, chief scientist for the Army Research Office.

Everitt manages one of the nation's largest programs in quantum computing. Other key players are the Defense Advanced Research Projects Agency, the National Security Agency and the Advanced Research and Development Activity.

Twice a year, since 1999, Everitt has chaired a meeting with the dozen or so federal agencies currently researching quantum computing. "My collaborative agencies are asking if one could be built," Everitt said in an interview. "Frankly, I have my doubts."

Everitt's effort is aimed at determining if the laws of physics would allow a quantum computer to exist. On that front, at least, things look hopeful. "So far the physics seem to be accommodating," Everitt noted.

If quantum computing seems esoteric, that's because it makes little sense to those who aren't particle physicists. Quantum mechanics is the study of matter at the atomic level, where the "normal" rules of physics don't apply.

Traditional computers encode information in a series of bits, and manipulate them in a specific order. They can do these manipulations very quickly. But a quantum computer could do a lot more at the same time, taking advantage of the oddities found at the atomic level.

Quantum objects can exist in multiple states simultaneously. By taking advantage of this phenomenon, quantum computers could perform simultaneous Factoring. Bringing that theory to a workable reality, however, is still a daunting challenge.

One problem facing scientists is the inability to keep the inner workings of a quantum computer stable. Quantum bits are like little tops spinning about an externally applied magnetic field at a known frequency (for example, 500 million revolutions a second.) If the spin rate changes even slightly, the bits' orientation can change and cause errors.

The premier application of a quantum computer lies in the field of encryption. But other benefits of even a selectively deployed quantum computer system could revolutionize the way military commanders allocate resources and execute operations.

The problem is often called the "Traveling Salesman," which tries to find the shortest route between X number of cities the vendor must visit. As the value of X increases, the problem gets exponentially harder. Finding the best route without testing every possible route against each other, one by one, and assessing results, is the holy grail of information technology, Everitt said.

Peter Shor, a researcher at AT&T's Bell Laboratories and now at MIT, devised the first quantum computer algorithm in the 1990s.

The ambition for quantum computing reached a near-fever pitch with the discovery of Grover's Algorithm--a powerful database search for a quantum computer that could find a needle in a haystack by identifying every strand of hay at once, rather than looking for the anomaly one-at-a-time. But Everitt said results have been disappointing so far. "We don't have an algorithm for that," he said. "There are whole classes of problems quantum computing can't be used for."

Everitt holds workshops to investigate other possible applications of the nascent technology. One such workshop in December was dedicated to the potential use of quantum computers in signal and image processing. "The answer was, 'Eh? We don't know."

That is the conundrum for quantum computers; the emergence of such a powerful tool comes with relatively few real-world applications. "Most people would never see or use a quantum computer," said Everitt. "This is not one of those things which will replace all computers."

Others are more optimistic. David Yepez, senior physicist at the Air Force Research Laboratory at Hanscom Air Force Base, Mass., sees great potential of quantum computers to force a massive change in information technology.

"If we are successful at nanometer scale engineering and in turn able to construct a large quantum computer, we can safely predict today that this ... will usher in another revolution tomorrow, which may be called the quantum revolution," he wrote in response to questions from National Defense.

In February 2004, Yepez wrote a memorandum to the Air Force on the topic of potential applications of quantum computing. Among the uses were drastic improvements in control, communications, intelligence, surveillance and reconnaissance.

Quantum clock synchronization could change operating procedures across the entire military, and using low-power quantum devices on space-based platforms would help the United States maintain space dominance via space-based quantum sensors, secure communications and on-board data processing.

Yepez acknowledges that large amount of work is needed to achieve these results.

Everitt said that in many ways, the Defense Department is waiting to see whether a quantum computer will become a reality before taking any serious steps towards fully perusing the technology. "I wouldn't say they're being myopic. I'd say they're being cautious," Everitt said. "They're betting on failure, and they may be right. We'll see soon enough."

One area where quantum mechanics has enormous potential in the defense market is for secure communications. Quantum cryptography enables parties to secure their communications by encoding a secret key with single photons, making any attempt at eavesdropping or code-cracking immediately discovered. The mere act of intercepting a photon to observe, read, copy or clone it irretrievably changes that photon. The more information that is obtained, the more photons are disrupted, and the more obvious the intrusion.

That is the basis of a project undertaken at Los Alamos National Laboratory. Researchers there, working with an international team, have developed a quantum cryptography technology. They received a $1.3 million prize from the European Union for their work on establishing quantum communications systems that run on particles of light.

The feasibility of quantum cryptography has been demonstrated in laboratories since the 1980s. Recent attempts use dedicated fiber-optic lines. The technology has problems, mostly related to imperfections of the photon sources.

There are already products on the market using quantum mechanics. A company called MagiQ Technologies is researching quantum computers, but also is selling quantum-based cryptography and communications systems. Quantum cryptographic key is already for sale, using the photon-by-photon encryption. Company officials said the initial markets for these quantum encryption keys are government agencies and financial institutions.

The flip side to the cryptography is code-breaking, an area where the U.S. military could find itself on the wrong end of quantum physics. Many U.S. government agencies employ today's premier encryption code, called RSA. The advent of a computer that could crack that code would be a grave threat, officials said.

But if a quantum computer of significant size and stability were created, Everitt said the reaction from the defense and intelligence communities would be immediate.

Quantum computers could process algorithms that the fastest existing supercomputer would need years to complete. That puts all codes and communications at risk--and could fuel the next boom in quantum computing research, Everitt said.