UCLA

The rapid growth of Internet of things (IoT) including mobile phones, portable devices, and remote sensor network systems have imposed both conceptually and technically new challenges. Among them, the most demanding requirements for the widespread realization of many IoT visions are security and low power. In terms of security, IoT applications include tasks that are rarely addressed before such as trusted sensing, secure computation and communication, privacy, and data right management. These tasks ask for new and better techniques for the protection of hardware, software, and data.

On the other hand, most of the IoT systems suffer from the problem of limited power sources, which in turn require the security on IoT devices to be lightweight. While low energy design is crucial for the successful deployment of resource-constrained IoT devices, their often physically accessible nature, as well as low energy budget restriction, have also contributed to rendering traditional cryptographic approaches insufficient to address all the security concerns.

In this dissertation, we present hardware-oriented security designs and synthesis techniques with the aim to reduce the system energy overhead while maintaining the security and reliability. We first demonstrate our work to analyze and enhance the properties of hardware security primitives. We emphasize on physical unclonable functions (PUFs) and use them to enable a wide range of applications, including private/public key communication, authentication, and multi-party communication. We then present a unique system reliability design with the use of non-volatile memory (NVM) to create a fault-tolerance application specific architecture with almost no timing overhead and low energy overhead. Finally, we demonstrate novel energy reduction and synthesis techniques applied on integrated circuit subsystems of IoT applications. The techniques we have applied and improved include near-threshold computing, dual-supply voltage optimization, and pipelining.