Skip to main navigation Skip to search Skip to main content

SaTC: STARSS: Small: Collaborative: Managing Hardware Security in Three-Dimensional Integrated Circuits

Project: Research

Project Details

Description

Vertically stacked three-dimensional (3D) integration of semiconductor chips is an emerging technology to ensure continued growth in transistor density and performance of integrated circuits (ICs). Despite the well-characterized advantages and limitations, the hardware security of such circuits has not received much attention. With shrinking number of trusted circuit manufacturers, trustworthiness of electronic devices is a growing concern. Vertical integration brings unexplored and unique challenges in managing hardware security. This research develops a cross-layer framework to manage and guarantee hardware security in 3D ICs. The systematic framework developed by this research assesses and enhances trustworthiness of 3D ICs for both commercial and defense-specific applications. Proposed project also supports diverse educational and outreach programs with emphasis on underrepresented groups in both New Hampshire and Stony Brook, Long Island. Existing hardware security solutions are for 2D ICs manufactured by a trusted foundry. Unfortunately, new security problems arise when multiple dies from different vendors are stacked in a 3D IC. Proposed research targets these security weaknesses with an emphasis on the trustworthiness of vertical communications. Solution mechanisms to prevent attacks on the vertical communication path will be investigated to achieve secure die-to-die communication. A secure power delivery architecture will be designed to detect small hardware Trojans in any die within the stack. A switching noise analysis based scheme will be investigated to detect malicious behavior in vertical communication paths.
StatusFinished
Effective start/end date10/1/1709/30/21

Funding

  • National Science Foundation: $146,667.00

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.