Design of an SRAM-based physical unclonable function system

Safety has been a critical concern in traditional data protection schemes. Placing the secret key in non-volatile memory makes protecting the key difficult and expensive if not impossible. Physical Unclonable Function (PUF) is an emerging primitive based on Integrated Circuit (IC) technology for eff...

Full description

Saved in:
Bibliographic Details
Main Author: Liu, Chao Qun
Other Authors: Chang Chip Hong
Format: Final Year Project
Language:English
Published: 2015
Subjects:
Online Access:http://hdl.handle.net/10356/63841
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Safety has been a critical concern in traditional data protection schemes. Placing the secret key in non-volatile memory makes protecting the key difficult and expensive if not impossible. Physical Unclonable Function (PUF) is an emerging primitive based on Integrated Circuit (IC) technology for efficient identification and device authentication. It takes the advantage of the unpredictable disorder of IC devices to generate device specific keys and is powerful to address the above problem. Static Random Access Memory (SRAM), which is commonly used as data storage component, has now been developed into a PUF by exploiting the physical disordered properties of SRAM devices. An SRAM based PUF, which can be used as both normal memory and a PUF, can save the overheads of implementing the two functions on two separate chips. However, the desired conditions of two modes, large process variation for PUF mode and small process variation for memory mode, are contradictory. This problem should be carefully considered when designing the SRAM-based PUF. This report presents a statistical analysis on SRAM failures and an insight on how to qualify the two modes concurrently. For supplement of the cell stability, three post-silicon techniques were compared and adopted to reduce the cell failure probability. Finally, the optimized design methods considering other constraints such as power and area will be proposed.