K
Kathleen Martin
Guest
The Internet of Things (IoT) refers to the growing number and type of devices that are connected to the internet. IoT devices range from Amazon Alexa and Google Home to bluetooth coffee makers and toothbrushes. Unlike cell phones and computers, these IoT devices may not have robust security. Hackers can infiltrate a network through a weaker device to access private information on devices with better security.
Creating security for IoT devices is challenging because they must be physically small (to fit in the device) and require relatively low power (to avoid draining the device). Usually, these devices are also not capable of performing complex computations. Carnegie Mellon University researchers at the Energy-Efficient Circuits and Systems (EECS) Lab from the Department of Electrical and Computer Engineering are inventing new ways to meet these challenges. In two recently published papers, Assistant Professor Vanessa Chen and her Ph.D. students Yuyi Shen and Jiachen Xu explored using variations from the manufacturing process for more robust security.
Within a network, the devices—also called nodes—already trust one another; that is, they openly communicate by sending and receiving signals. If a foreign device tried to interact with a node, the network would ignore it. One way that hackers gain access to networks is by impersonating a trusted node. Thus creating security measures must involve a robust process for verifying the identity of any device attempting to communicate.
"The typical approach for implementing a wireless security system is through some kind of cryptographic mechanism, which requires computational power that some IoT devices have trouble supporting," Shen said. "The focus of our research is on one specific form of low power security mechanism called radio frequency fingerprinting."
Radio frequency fingerprinting (RFF) refers to a method of identifying devices by exploiting hardware variations that arise despite the precision used during the manufacturing process. These variations result in unique features in the radio waves the device transmits. After signal processing, these features can be used to identify a specific device.
One way to make RFF harder for hackers to identify, and thus mimic, is to change the fingerprint's features. This is a non-trivial task, especially given the fact that the RFFs result from unintentional manufacturing deviations. One of the researchers' papers looked at using power amplifiers to change a device's signal features.
"Usually, each device will have fixed hardware characteristics that might change with the environment or slowly over time," Xu said. "But this power amplifier is capable of reconfiguring itself to generate various radio frequency fingerprints in one device, preventing people attacking the device from mimicking the hardware characteristics."
Continue reading: https://techxplore.com/news/2022-02-fingerprinting-internet.html
Creating security for IoT devices is challenging because they must be physically small (to fit in the device) and require relatively low power (to avoid draining the device). Usually, these devices are also not capable of performing complex computations. Carnegie Mellon University researchers at the Energy-Efficient Circuits and Systems (EECS) Lab from the Department of Electrical and Computer Engineering are inventing new ways to meet these challenges. In two recently published papers, Assistant Professor Vanessa Chen and her Ph.D. students Yuyi Shen and Jiachen Xu explored using variations from the manufacturing process for more robust security.
Within a network, the devices—also called nodes—already trust one another; that is, they openly communicate by sending and receiving signals. If a foreign device tried to interact with a node, the network would ignore it. One way that hackers gain access to networks is by impersonating a trusted node. Thus creating security measures must involve a robust process for verifying the identity of any device attempting to communicate.
"The typical approach for implementing a wireless security system is through some kind of cryptographic mechanism, which requires computational power that some IoT devices have trouble supporting," Shen said. "The focus of our research is on one specific form of low power security mechanism called radio frequency fingerprinting."
Radio frequency fingerprinting (RFF) refers to a method of identifying devices by exploiting hardware variations that arise despite the precision used during the manufacturing process. These variations result in unique features in the radio waves the device transmits. After signal processing, these features can be used to identify a specific device.
One way to make RFF harder for hackers to identify, and thus mimic, is to change the fingerprint's features. This is a non-trivial task, especially given the fact that the RFFs result from unintentional manufacturing deviations. One of the researchers' papers looked at using power amplifiers to change a device's signal features.
"Usually, each device will have fixed hardware characteristics that might change with the environment or slowly over time," Xu said. "But this power amplifier is capable of reconfiguring itself to generate various radio frequency fingerprints in one device, preventing people attacking the device from mimicking the hardware characteristics."
Continue reading: https://techxplore.com/news/2022-02-fingerprinting-internet.html
