报告人简介：
Chun Jason Xue香港城市大学计算机科学系副教授，主持完成了包括香港政府科研计划及公司合作在内的20余个科研项目，参与完成了国家863计划项目的研究工作。在国际顶级学术会议和期刊上发表论文200余篇，包括存储领域顶级会议FAST，调度算法领域顶级会议RTSS，体系结构领域顶级期刊TC、TCAD，数据处理领域顶级期刊TKDE等。担任Transactions on Storage (TOS)，Transactions on CPS等多个国际期刊的副主编，担任DAC，DATE等多个国际会议的审稿人。主要研究方向包括移动与嵌入式体系结构、非易失性存储架构、存储与大数据、物联网等。
报告摘要：
Mobile devices, such as smartphones, have become a necessity in our daily life. NAND flash memory is the primary choice of data storage for mobile devices due to its high performance, shock resistance and low power consumption. However, compared to high-end solid-state drives, mobile flash storage does not have the luxury of sophisticated hardware and firmware features because of the resource-constraint limits which make it a challenge to apply traditional techniques for I/O performance improvements. This talk is about the I/O system optimization dedicating to mobile environments.
First, this talk will introduce a comprehensive study of file fragmentation on mobile devices. Users may notice that after being used for a long time, mobile devices begin to exhibit sluggish response. This study takes a three-step approach: Firstly, this study designed a set of reproducible file-system aging processes based on User-Interface (UI) script replay. Through the aging processes, it confirmed that file fragmentation quickly emerged. Secondly, based on the workloads of a selection of popular mobile applications, this study observed that file fragmentation did impact user-perceived latencies. Thirdly, this study evaluated existing treatments of file fragmentation, including space preallocation, conventional file defragmentation, and a state-of-the-art defragmenter to understand their efficacies and limitations.
Second, this talk will present a lightweight data compression technique at the flash controller to reduce write pressure on mobile flash storage. It first characterizes data compressibility based on real smartphones, and the analysis shows that write traffics bound to mobile storage volumes are highly compressible. This technique is the first to investigate firmware-based data compression for mobile flash storage without adding extra data compression hardware. Experimental results demonstrate the proposed techniques outperform state-of-the-art schemes in terms of.
