Enabling Reliable and Efficient Computing for Energy Harvesting Powered IoT Devices
Date: Tue, April 02, 2019
Time: 10:00am
Location: Holmes 389
Speaker: Mimi Xie, University of Pittsburgh
Abstract:
The emerging era of Internet-of-Things (IoT) has raised fundamental challenges in powering billions of devices. Energy harvesting (EH) techniques, which scavenge energy from ambient environment, have become a promising and favorable long-term solution to address this challenge. While EH has great potential, it has its own drawback - intermittency, which can interrupt the systems frequently and result in execution progress setback. To overcome this drawback, non-volatile processors (NVPs) equipped with emerging non-volatile memories have been developed to save the intermediate computation state before the power outage so that the computation can be resumed after power comes back. While NVPs can address the power intermittency problem, it introduces several new challenges. Without new hardware and software support, the program execution resumed from the last checkpoint might not execute correctly and causes inconsistency problem. Frequent and large backup overhead will also degrade the performance. This talk will discuss the key challenges of modern intermittently powered embedded system design in both software and hardware levels and present efficient both complier and architecture optimizations to enable reliable and efficient embedded systems for EH-powered IoT devices. In the end, the design trends for the next-generation IoT system and future research directions will also be discussed.
Bio:
Mimi Xie is currently a PhD candidate in the Department of Electrical and Computer Engineering at the University of Pittsburgh. Her research interests span across emerging non-volatile memory technology, intermittently powered system, and memory security, with an emphasis on software and hardware co-design of reliable, secure, and energy efficient embedded systems and their applications in IoT. Her research is driven by both emerging memory technologies and modern IoT applications. She is a recipient of Women Faculty Research Award and A. Richard Newton Young Student Fellowship.