Kimble possesses a courtly Texas gentleman’s manner, as I discovered after his lab manager found him 15 minutes on the schedule following two weeks when the physicist was away, making presentations at four conferences on two continents. Those 15 minutes became a tutorial on recent technical advances in verifying and quantifying entanglement. Measurement is the central problem in quantum mechanics, since any particle or system exists in a quantum state only until another system, whether one as slight as a stray air molecule or as complex as a human observer, gains information about it and thereby collapses that state. This is mind-bendingly abstruse stuff. Aside from discussing quantum metrology, though, Kimble made one easily graspable assertion: “Our society’s technical base is information commerce. In the next 20 years, quantum information science–a fusion of computer science and quantum mechanics that didn’t exist 20 years ago–will radically change that commerce.”
The revolutionary technology that Kimble envisions is large quantum networks, resembling the Internet but relying on entanglement. What inherent advantages would promote the development and adoption of such networks?
Substantial ones. Quantum networks have already been built on a limited scale. In 2004, the world’s first permanent quantum cryptography system was activated in Cambridge, MA, linking Harvard, Boston University, and DARPA contractor BBN Technologies (formerly known as Bolt Beranek and Newman, under which name the company created the original ARPAnet). Today, id Quantique, a Swiss company, and MagiQ Technologies, a U.S. one, offer commercial modules using optical fiber to transmit quantum keys, in the form of photons encoded as bits by controlling their polarization, over limited distances that top out at about 100 kilometers. Since attempted interception of these light particles would disturb their state and expose eavesdropping, such quantum cryptography systems offer absolute data security.
Furthermore, the prospect of quantum computing was what provided the initial impetus for research into quantum networks. If such computing can be done seriously (so far, experiments have used at most seven qubits, or quantum binary digits), it promises to surpass classical computing in significant respects. Scott Aaronson, an MIT expert on computational complexity, cites the algorithm published in 1994 by MIT mathematician Peter Shor as the breakthrough that proved quantum computing a viable proposition by demonstrating that it could factor very large numbers in reasonable computing time. Because that task has been beyond classical computers, most public-key cryptography has hitherto been based on factoring large numbers. But it would be vulnerable to cryptanalysis based on quantum computing. As Aaronson says, “That’s why the National Security Agency is interested in quantum computing.” Quantum cryptography, however, would offer data security against quantum code-breaking as well as against regular cryptanalysis.
Besides ensuring the security of data, the quantum wide-area repeater networks, or QWANs, that Kimble has in mind would possess few of current networks’ latency issues–indeed, could be as nearly instantaneous as light speed allows. Moreover, the exponential parallelism that would give quantum computing its power–with two entangled particles, or qubits, representing four different values, four qubits 16 values, and so on–ought to apply to networks of quantum computing devices. Kimble says, “Though there’ll be a largest size attainable for the state space of individual quantum processing units, it’ll be possible to surpass that by linking those units together into a fully quantum network.” A quantum computer’s “state space” is the full range of potential states in which the computer could exist. When a quantum algorithm is run, this computational process collapses that state space and shrinks the computer’s range of possible states down to a single one: the correct answer to the given problem. With a network of quantum computers, Kimble is claiming, the exponential computational power of each device would be multiplied exponentially.