Sign In

Communications of the ACM

ACM News

Software to Combat Hacking Technique Wins Award

View as: Print Mobile App Share:
Virginia Tech assistant professor Daphne Yao

"Keystroke dynamics is an inexpensive biometric mechanism that has been proven accurate in distinguishing individuals," says Virginia Tech assistant professor of computer science Daphne Yao.

Credit: Virginia Tech

One of the serious threats to a user's computer is a software program that might cause unwanted keystroke sequences to occur in order to hack someone's identity. This form of an attack is increasing, infecting enterprise and personal computers, and caused by "organized malicious botnets," says Daphne Yao, assistant professor of computer science at Virginia Tech.

To combat the "spoofing attacks," Yao and her former student, Deian Stefan, now a graduate student in the computer science department at Stanford University, developed an authentication framework called "Telling Human and Bot Apart" (TUBA), a remote biometrics system based on keystroke-dynamics information.

Their work won a best paper award at CollaborateCom 2010, the 6th International Conference on Collaborative Computing, held in Chicago in October 2010.

Yao holds a patent on her human-behavior driven malware detection technology, including this keystroke anti-spoofing technique. Her technology for PC security is currently being transferred to a company. The license agreement between the company, Rutgers University (Yao's former institution), and Virginia Tech is expected to be finalized in the coming weeks.

Internet bots are often described as web robots. They act as software applications that run automated tasks over the Internet. Bots usually perform simple and repetitive tasks, but at a much higher rate than would be possible for a human alone. When used for malicious purposes they are described as malware.

"Keystroke dynamics is an inexpensive biometric mechanism that has been proven accurate in distinguishing individuals," Yao says, and most researchers working with keystroke dynamics have focused previously on an attacker being a person.

The uniqueness of Yao and Stefan's research is its focus on identifying when a computer program designed by a hacker was producing keystroke sequences "in order to spoof others," they say. Then they created TUBA to monitor a user's typing patterns.

Using TUBA, Yao and Stefan tested the keystroke dynamics of 20 individuals, and used the results as a way to authenticate who might be using a computer.

"Our work shows that keystroke dynamics is robust against the synthetic forgery attacks studied, where the attacker draws statistical samples from a pool of available keystroke datasets other than the target," Yao says.

Yao and Stefan also describe in their paper, "Keystroke-Dynamics Authentication Against Synthetic Forgeries," how keystroke dynamics can be used "as a tool to identify anomalous activities on a personal computer including activities that can be due to malicious software."

In January of 2010, Yao won a prestigious Faculty Early Career Development (CAREER) grant from the U.S. National Science Foundation to develop software that differentiates human-user computer interaction from that of malware. The five-year $530,000 grant is funding Yao's research for building a new malicious software detection system for personal computers that is able to accurately differentiate network behaviors of a legitimate human user from a malware program.


No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account