PUF-Based Unique ID Generation in IoT Devices
With the advent of the IoT era, there is an increasing demand for risk control of all types of IoT devices, one of the bases of which is the unique identification of the device. Pure software-based approaches are not tamper-resistant and 'clonable'. On the other hand, adding a secure element chip is too expensive for a small IoT device.
PUF, or physical unclonable function, is one of the most reliable, well-tested approaches with a lot of benefits such as tamper-resistant, unclonable, high information entropy, and the characteristics of which are guaranteed in the process of chip manufacturing.
With academic progress in this area for a decade, a foreseeable large-scale commercial deployment is just around the corner. But on the engineering side, there are still many problems waiting to be solved. For instance, existing schemes rely on modifications to the bootloader, which makes hot updates and deployments very difficult. Furthermore, due to the fragmentation status of IoT, too many chips, architectures, bootloaders, operating systems are required to be adapted, which makes the deployment procedure even more difficult.
In addition, with the emergence of new types of RAM such as RRAM, there is also the possibility to apply PUF migration of traditional PUF technology to new types of electronic devices.
- Optimize the existing PUF approaches for large scale deployment on different operating systems and architectures.
- Construct new ID generation algorithms based on existing PUF datasets, meet stability requirements.
- Find new data (entropy) sources in the scene of IoT, meeting the requirement of ID generation algorithms.
- Expected deliverables: intellectual properties, and a technical report.
Related Research Topic
- Clone/tamper detection of IoT devices
- DRAM-based physical unclonable function
- CCD-based uniqueness source, or PRNU
- The new sources of entropy/uniqueness in embedded systems
- Biometric authentication data processing
Suggested Collaboration Method
AIR (Alibaba Innovative Research), one-year collaboration project.