University research aims at more secure wireless network

Researchers at Carleton University, Ottawa, Canada, have reported positive results for a novel means of securing Wi-Fi and other wireless networks from hackers and other unauthorized intrusion.
The technology depends on the RF signal "fingerprints" or profiles that make every wireless transceiver in the world virtually unique. The RF fingerprints are the result of variations in the silicon and other electronic components that comprise the transceiver.

Although the components all fall within the manufacturing tolerances required by the vendor and generate valid signals, the combinations of their variances create unique signal characteristics, says Jeyanthi Hall, a graduate student at the university who is the lead researcher for the project supervised by professors Michel Barbeau and Evangelos Kranakis.

Variances are most evident in the transient signals created when the transceiver attempts to gain access to the network. In a Wi-Fi network, this means the fingerprint is acquired in approximately 2 microseconds.
A probabilistic neural network is used to compare the fingerprint to others stored in the access point (or some central location in the network) that have been verified by the network system administrator as authentic.


The researchers are also exploring the use of self-organizing map technology and clustering technology to reduce the storage capacity required for the authenticated signatures and to speed authentication.
Algorithms from The MathWorks.com MATLAB technical computing software are tuned and used for the authentication process. During the research phase of the project, the transient RF signals from the transceivers are acquired using Anritsu's Signature High Performance Signal Analyzer.

As the technology moves into more refined stages, Hall said, the signal analyzer will be replaced by a DSP-based data acquisition board.

The signal fingerprinting technology being researched at Carleton University complements and utilizes traditional security measures such as MAC-address control lists.

With spoofing techniques, hackers can circumvent the effectiveness of a MAC-address control list. With RF fingerprinting included in the security arrangements, however, a transceiver that dishonestly reports itself as having a specific MAC address can be uncovered by checking its fingerprint against the authenticated transceiver's.

Hall's research still has several hurdles to clear before it can appear as commercial product. Chief among them are scalability and the stability of the algorithms employed to create the fingerprint and compare it to other RF fingerprints.