Associate Research Fellow  |  Chang, Yuan-Hao  
Personal (New window)
Research Descriptions

My research area is mainly on embedded systems, with a prime focus on improving the performance, reliability, and endurance of their storage systems with different types of non-volatile memories. It later grows to consider byte-addressable non-volatile memory as both main memory and storage of embedded systems. 

Embedded systems, especially battery-powered consumer electronics and mobile computing systems such as smartphones, usually adopt flash storage devices as their storage systems. However, due to the shrinking of the fabrication process and the advances of manufacturing technology, the density and capacity of flash-memory chips grow dramatically in recent years, but performance, reliability, and endurance have become major design issues in related product designs. To address the above issues, we are interested in the solution at the storage device level. Our research goal is to resolve these issues with (flash) management designs in the software/firmware layer (i.e., flash translation layer) of flash storage devices, where the flash translation layer (FTL) is mainly in charge of address translation, garbage collection, and wear leveling for the underlying flash chips.

Another recent research focus is on the operating system designs for embedded systems with one-memory architecture, which is the architecture with byte-addressable non-volatile memory (NVM) as both main memory and storage of embedded systems. To exploit the benefits of the one-memory architecture, conventional operating systems need to be modified because they manage main memory and file system separately, thus resulting in many unnecessary data movements between main memory and storage when both main memory and storage are in the same address space under the one-memory architecture. Thus, our research goal is to propose new operating system designs, including new memory management and file systems designs, that can jointly manage memory andstorage space on the same NVM device, so as to fully utilize the advantages brought by the one-memory architecture.