Developing Ultra-Low Power Circuits to Prevent Denial of Sleep IoT Attacks
ECE Assistant Professor Aatmesh Shrivastava received a $356K NSF grant for “Energy and Activity Analysis-based On-chip Methods for Mitigating Denial-of-Sleep Attacks in Ultra-low Power IoT Devices.”
Abstract Source: NSF
The finite and often scarce available sources of energy (e.g., batteries) in low-power internet-of-things (IoT) sensing devices have presented themselves as unique weak points often being vulnerable to cyber-attacks. This project aims to develop an innovative hardware security mechanism to protect the available energy source to support longer operational life-time for such devices. This project will develop a local, on-chip hardware security mechanism to enhance security at the root but with lower power and cost overhead. The proposed techniques are transformative and has the potential to harden security in ultra-low power applications. Research outcomes from this project will be disseminated as tutorials at leading hardware security and circuit conferences to raise awareness about security vulnerabilities in low-power sensing devices. This project will provide a platform for training graduate and undergraduate students on circuit design, sensing systems, and cybersecurity. New educational materials for the general public on the security applications of circuit design considerations and possible tradeoffs for security, area, and power will also be developed. Engagements of undergraduate students in research through the research experience for undergraduates (REU) program and annual summer projects are planned for visiting students from minority serving institutions, and high schools.
This research aims to develop ultra-low power, on-chip, hardware security solutions for IoT sensing devices. Low-power sensing devices duty-cycle their communication and sensing activity and as such spend the large fraction of their time sleeping to conserve and harvest energy. Denial-of-sleep attacks target this feature to deny them of much-needed sleep by initiating frequent wake up and repeated communication requests to quickly drain their stored energy. The goal of this project is to explore new methods of defense against denial-of-sleep attacks by analyzing the energy consumption patterns of a sensing device. Their activity and energy consumption patterns can be extracted as features and learned to differentiate a denial-of-sleep attack from a regular activity of the device. This project will develop an on-chip learning method to prevent denial-of-sleep attacks. An energy monitoring system will be developed to continuously track the availability and consumption of the energy in a sensing device. The monitoring system will tap into key circuit points of the power management system to realize a lower cost hardware solution. Further, the architecture of a received signal strength indicator (RSSI) circuit will be advanced to operate at ultra-low power level with a lower area overhead. This circuit will reside outside of the radio front-end and will continuously monitor the RF channel for jamming attacks. The proposed method allows ultra-low power consumption and lower area which ensures lower cost overhead as new hardware security solutions to be translational across a variety of IoT sensing applications.
This award reflects NSF’s statutory mission and has been deemed worthy of support through evaluation using the Foundation’s intellectual merit and broader impacts review criteria.