Research Area
High-performance computing and Data Center
Cloud infrastructure and orchestration
Big data processing and management
Database (RDBMS, Key-Value, NoSQL)
Operating System, especially file systems (EXT4, F2FS, XFS)
Next-generation storage (Smart SSD, ZNS, FDP)
Large-scale simulation (Quantum computing simulation, particle simulation, earth simulation)
IoT and Edge Device
Autonomous system management in a large scale supercomputer
빅데이터 처리 시스템 (슈퍼컴퓨터, 클라우드, 데이터센터)에서의 자율적 시스템 최적화
As a large-scale computing system, data centers, and cloud systems are rapidly evolving. How will we allocate computing resources and manage complex software/hardware layers? Our goal is to analyze the system and optimize system resource allocation.
슈퍼컴퓨터, 클라우드, 데이터센터 사이의 경계는 희미해지고 계속해서 발전하고 있습니다. 복잡한 하드웨어 및 소프트웨어 환경에서 시스템을 어떻게 관리해야 할까요? 사용자를 위해 시스템을 분석하고 시스템 자원을 최적화하려합니다.
Previous Works
Big data analysis of a large-scale production HPC system (Cori supercomputer which debuted as Top 6th in Top 500 list)
Found strong correlation using various correlation analysis algorithms
Proposed a prediction scheme using machine learning approaches such as random forest and CNN
Big data analysis of a complex HPC system using large scale logs
In-depth correlation study with visualization
Performance prediction model and prediction results
Next-generation Computing architecture
Emerging hardware: ZNS SSD and CXL, and their adaptation to big data processing systems.
Heterogenous system and Multi-GPGPU scheduling
Burst-Buffer and next-generation storage devices (NVDIMM, NVMe, and tiered storage system)
Access pattern-based performance prediction and optimization (LSTM, CNN, and more)
System optimization for massively parallel and distributed systems (10,000+ cores and exascale computing)
Blockchain
KV database for large-scale blockchain system
Smart contract and block-aware performance optimization
Next-generation blockchain-based storage (IPFS)