Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

University of Toronto

$37,000.00

Automated Testing Techniques for Hardware Security and Reliability

FY 2023-24

Summary
In recent years, critical security vulnerabilities (like Spectre and Rowhammer) have been discovered in processor designs, affecting billions of computers. These provide malicious software new pathways to leak sensitive data and tamper data in systems. Moreover, with the end of Moore's law, new reliability challenges are emerging. Companies like Google and Meta, operating fleets of hundreds of millions of CPU cores, recently reported a dramatic increase in crashes due to silent data corruptions in CPUs. Thus, security and reliability have become a first-order concern for hardware of this decade and beyond. Unfortunately, we lack robust automated techniques to detect such issues. On the security front, a series of hardware defenses have been broken by newer attacks, spurring an arms race. On the reliability front, the vulnerability to faults in DRAM and CPUs increases with each generation. To detect and tolerate such increasing vulnerability levels, this project develops automated testing techniques for hardware security and reliability along three thrusts: Thrust-1: Automated Testing for Hardware Side Channels at Design-Time. Typically, new side-channel attacks (like new Spectre variants) are detected in commercial hardware only after it is designed and deployed. Subsequently, it takes years to fabricate and deploy new hardware defenses, rendering systems perpetually insecure. To address this, we develop techniques to test the security of a hardware defense at design time against adaptive side channels. We enable this by fuzz-testing processor designs in architectural simulators. This technique can allow vulnerabilities to be detected and addressed at hardware design time, thus ending the arms race in hardware side channels. Thurst-2: Durable Defenses against Rowhammer Attacks in DRAM. Given the increasing Rowhammer vulnerability in each DRAM generation, this project will study the Rowhammer vulnerability for emerging DDR5 devices, using programmable FPGA-based memory-controllers and commercial systems. Subsequently, this work will explore the potential of adaptive attacks against emerging academic defenses as vulnerability increases, and develop solutions to safeguard DRAM integrity regardless of increasing vulnerability levels. Thurst-3: Automated Testing for Faulty CPUs at Run-Time. Fleet-wide testing of CPUs in data-centers consumes considerable power. Consequently, this work proposes to improve both testing coverage and efficiency with new CPU interfaces that can expose micro-architectural state information to software during the testing process. With better runtime feedback to the test-generation algorithms, this research will help increase the micro-architectural coverage of testing tools such as SiliFuzz, while running fewer tests overall to improve efficiency. Overall, the techniques developed in this project will enable more durable security and reliability for future computing hardware.
Program
Discovery Grants Program - Individual
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
RGPIN-2023-04796
Name nom
Saileshwar, Gururaj
Department département
Department of Mathematical and Computational Sciences, UTM
Organization id
31
Institution établissement
University of Toronto
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2023
Competition year année de concours
2023
Award amount
37000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1510
Committee name en
Electrical and Computer Engineering
Committee name fr
Génie électrique et informatique
Area of application code
513
Area of application group en
Manufacturing processes and products
Area of application group fr
Procédés et produits industriels
Area of application en
Electrical and electronic machinery and equipment (including computer hardware)
Area of application fr
Machinerie et équipement électriques et électroniques (y compris matériel informatique)
Research subject code
2701
Research subject group en
Information technology
Research subject group fr
Computer hardware
Research subject en
Technologie de l'information
Research subject fr
Matériel informatique
Application title
Automated Testing Techniques for Hardware Security and Reliability
Keyword
Automated Testing Techniques; Computer Architecture; Fuzzing; Hardware Security; Side-Channel Attacks; Silent Data Corruption; System Security; Transient Execution Attacks
Application summary
In recent years, critical security vulnerabilities (like Spectre and Rowhammer) have been discovered in processor designs, affecting billions of computers. These provide malicious software new pathways to leak sensitive data and tamper data in systems. Moreover, with the end of Moore's law, new reliability challenges are emerging. Companies like Google and Meta, operating fleets of hundreds of millions of CPU cores, recently reported a dramatic increase in crashes due to silent data corruptions in CPUs. Thus, security and reliability have become a first-order concern for hardware of this decade and beyond. Unfortunately, we lack robust automated techniques to detect such issues. On the security front, a series of hardware defenses have been broken by newer attacks, spurring an arms race. On the reliability front, the vulnerability to faults in DRAM and CPUs increases with each generation. To detect and tolerate such increasing vulnerability levels, this project develops automated testing techniques for hardware security and reliability along three thrusts: Thrust-1: Automated Testing for Hardware Side Channels at Design-Time. Typically, new side-channel attacks (like new Spectre variants) are detected in commercial hardware only after it is designed and deployed. Subsequently, it takes years to fabricate and deploy new hardware defenses, rendering systems perpetually insecure. To address this, we develop techniques to test the security of a hardware defense at design time against adaptive side channels. We enable this by fuzz-testing processor designs in architectural simulators. This technique can allow vulnerabilities to be detected and addressed at hardware design time, thus ending the arms race in hardware side channels. Thurst-2: Durable Defenses against Rowhammer Attacks in DRAM. Given the increasing Rowhammer vulnerability in each DRAM generation, this project will study the Rowhammer vulnerability for emerging DDR5 devices, using programmable FPGA-based memory-controllers and commercial systems. Subsequently, this work will explore the potential of adaptive attacks against emerging academic defenses as vulnerability increases, and develop solutions to safeguard DRAM integrity regardless of increasing vulnerability levels. Thurst-3: Automated Testing for Faulty CPUs at Run-Time. Fleet-wide testing of CPUs in data-centers consumes considerable power. Consequently, this work proposes to improve both testing coverage and efficiency with new CPU interfaces that can expose micro-architectural state information to software during the testing process. With better runtime feedback to the test-generation algorithms, this research will help increase the micro-architectural coverage of testing tools such as SiliFuzz, while running fewer tests overall to improve efficiency. Overall, the techniques developed in this project will enable more durable security and reliability for future computing hardware.