Research Interests
Distributed Storage and Computing
With the advancement of mobile, IoT, and cloud technologies, a large amount of data is being generated, stored and analyzed. Distributed storage and computing technologies play a vital role in data storage and management. We are conducting distributed file system research for cloud storage systems. Recently, we are studying deduplication techniques to remove redundant data in distributed shared-nothing storage systems. We are also working on integrating hardware accelerators such as GPU and I/O scheduling algorithms to speed up the deduplication operations.
Operating System
Non-volatile memory (NVM) is byte-addressable, so it can be used as a main memory and storage replacement. The emergence of non-volatile memory in memory hierarchy has introduced new research directions in the operating system (OS) domain. We are conducting the NVM-based memory file system research for scalable I/O service of many-core servers. In addition, we are currently conducting research on key-value storage devices for unstructured data storage, and also studying communication protocol between hosts and devices, and firmware softwares.
Computer System Security
Attacks on data stored on individuals and on corporate computers are increasing worldwide by hacking and malware attacks. Since companies store important confidential data, including customer data, they must prevent unauthorized access to data and minimize damage from data tamper attacks. We are conducting research on system software-level defense techniques to protect against data tampering attacks such as Ransomware attacks.
|
