While quantum computers may one day be capable of unprecedented calculations, they are incredibly vulnerable to failure caused by unauthorized activity when networked together. Daniel Lidar, an associate professor in electrical engineering, chemistry, and physics at the University of Southern California, and Lian-Ao Wu, a research associate in the Chemical Physics Theory Group at the University of Toronto, are working on ways of defending quantum computer networks against something as small as a read access to a single qubit on one machine, which would require a network-wide reset. Their solution has been to only send messages at prearranged, seemingly random intervals, use long average wait times between legitimate network connections, and fill the rest of the network time with decoy transmissions. Performance advantages could be maintained while reducing the chance of a successful attack. “We would not want to use this method against any threat beside malware, because it is not efficient,” says Lidar. “We are talking the network down for a long period of time.” Quantum computers must be protected against “stray cosmic rays and things like that–if they interact with this stuff, then something changes and the computer crashes,” according to John Lowry, a principal scientist at the Internet service provider BBN, and member of the company’s research team working on the DARPA quantum network. “The thing is that people could do that on purpose.” Lidar doesn’t know what form an attack would come in; data destruction or circumventing a calculation would be the easiest. He says, “Quantum malware to us just looks like any malicious instruction sent to an attacker. As long as we can keep the local nodes free from malicious intruders and build a heavily fortified castle around them, we can assume the ancilla qubits are malware free.”
Source: ACM TechNews.
Click here to view full article on SecurityFocus.