Storage and Systems Research

image
High-performance computing depends on efficient, scalable storage subsystems. In this project area, I concentrate on designing and optimizing multi-tiered storage architectures—incorporating technologies like NVMe SSDs, burst buffers, and archival systems—to handle ever-growing data demands. My work includes building automated data-placement frameworks (e.g., Data Jockey) that relieve scientists from manual file management across multiple storage tiers, as well as performing OS-level and SSD firmware optimizations to boost I/O throughput.

Having contributed significant kernel-level enhancements, cross-layer SSD firmware modifications, and custom FPGA-based emulation platforms, I have a deep understanding of how to improve performance and reliability across the entire storage stack. These innovations pave the way for future HPC systems to handle large-scale simulation outputs, checkpoints, and analytics workloads more effectively and with lower latency.

Related Projects
Data Jockey - Multi-tiered Storage Systems for HPC centers (2016 - 2020)

Technology Integration Group, Oak Ridge National Laboratory, USA

Mar 1, 2017
NVM-e SSD Evaluations for Burst Buffers (2016 - 2017)

Technology Integration Group, Oak Ridge National Laboratory, USA

Mar 1, 2017
Enhanced OS & SSD Cross Layer I/O Optimization (Greybox SSDs) (2015 - 2017)

Distributed Computing Laboratory, Seoul National University, South Korea

Jan 1, 2015
Hardware & Software Co-optimization for High Performance / High Functionality SSDs in Data Centers (2014)

Research subcontracted from Samsung Electronics, Seoul National University, South Korea

Jan 1, 2014
OS Optimizations for Mulitple Flash SSDs (2013 – 2015)

Distributed Computing Laboratory, Seoul National University, South Korea

Jun 1, 2013
Optimized Systems for Big Data: Workload analysis for applying NVRAM components (2012 - 2013)

Internship, IBM Austin Research Laboratory, USA

Sep 8, 2012
Development of a File System based on Phase Change Memory to Support more than 200,000 IOPS for Peta-scale Computing (2011 - 2012)

Research Grant, Ministry of Knowledge Economy (MKE) of Republic of Korea, Seoul National University, South Korea

May 1, 2011
A Research on a Hierarchical Storage Architecture for Cloud Computing (2010 - 2012)

Research Grant, National Research Foundation of Korea (NRF), Seoul National University, South Korea

Aug 1, 2010