University of Massachusetts Amherst

Search Google Appliance

Links

UMass Electrical Engineers Describe Pioneering Hardware Security Device, Supported by Memristors, in Nature Communications

Qiangfei Xia

Qiangfei Xia

J. Joshua Yang

J. Joshua Yang

Professors Qiangfei Xia and J. Joshua Yang of our Electrical and Computer Engineering (ECE) Department led a team of scientists who have developed a groundbreaking new type of hardware security device enabled by memristors, or resistive switching devices, as described in an article in the prestigious scientific journal Nature Communications. The title of the new article is “A Novel True Random Number Generator Based on a Stochastic Diffusive Memristor.” This work paves the way for memristors in hardware security applications for the era of the Internet of Things (IoT).

As the research team says in its Nature Communication article, “The intrinsic variability of switching behavior in memristors has been a major obstacle to their adoption as the next generation universal memory. On the other hand, this natural stochasticity can be valuable for hardware security applications. Here we propose and demonstrate a novel true random number generator (TRNG) utilizing the stochastic delay time of threshold switching in an Ag:SiO2 [silver : silicon dioxide] diffusive memristor, which exhibits evident advantages in scalability, circuit complexity, and power consumption.”

The research team is composed of: Xia, Yang, Hao Jiang, Daniel Belkin, Siyan Lin, Zhongrui Wang, Yunning Li, Saumil Joshi, Rivu Midya, Can Li, and Mingyi Rao of the UMass Amherst ECE department; Sergey E. Savel’ev from the Department of Physics at Loughborough University in the UK; as well as Mark Barnell and Qing Wu of the Air Force Research Lab Information Directorate in Rome, New York. Co-first author Belkin was a freshman at Swarthmore College who came to work with Xia on this project last summer under the sponsorship of the Research Experiences for Undergraduates program of the National Science Foundation.

As the researchers explain the background for their project, IoT is a network of devices, sensors, and other items of various functionalities that interact and exchange data electronically. Because of the explosive growth in the number of IoT objects (estimated to be 50 billion by 2020) and overwhelming reliance on cyberspace, the existing hardware infrastructure is increasingly vulnerable to a wide range of security threats. Because software-based, data-securing methods are no longer sufficient because they are easily attacked, hardware security systems therefore become critical.

One security solution is a TRNG, a hardware component that generates a string of random bits which can be used as a cryptographic key. It relies on intrinsic stochasticity in physical variables as a source of randomness. However, all prior TRNGs have suffered from drawbacks in scalability, circuit complexity, or relied on post-processing such as a “Von Neumann corrector” to remove bias from the generated bit sequences.

But memristors have the capability of neutralizing these drawbacks. Memristors have been proposed and demonstrated for a broad spectrum of applications because of their attractive properties, such as low power consumption, fast switching speed, high endurance, great scalability, and CMOS-compatibility. The intrinsic variation in switching parameters is a major challenge for some applications such as non-volatile memory. However, this random behavior can be very useful in stochastic computing and hardware security applications.

“Here, we propose and demonstrate a novel TRNG based on a diffusive memristor, a newly developed volatile device that relies on the diffusion dynamics of metal atoms in the memristive layer,” as the Nature Communications article explains. “The device switches to a low resistance state under a voltage pulse after a random delay time, and relaxes back to the high resistance state spontaneously upon removal of the applied electrical bias. We used the intrinsic stochasticity of the delay time as the source of randomness and built a TRNG unit that consists only of a diffusive memristor, a comparator, an AND-gate, and a counter.”

According to the article, compared with previous TRNGs based on non-volatile memristors, the self-OFF-switching behavior in the diffusive memristor greatly reduces energy consumption, since no RESET process is required. In addition, this new TRNG also has evident advantages in circuit complexity “because the randomness is generated and harvested directly using simple elements. The diffusive memristor TRNG can easily be incorporated into memory subsystems, greatly increasing the security and the area efficiency.”

More importantly, the article adds, the bits generated by this diffusive memristor TRNG passed all the 15 randomness tests, without any post-processing, in Special Publication 800-22 of the National Institute of Standards and Technology. (August 2017)