Program

Oct 22 (Thu): TBD
Time	Subject	Speaker
13:40 - 14:00	Welcome Reception
	Session I (Session Chair: Donghyun Kang, Dongguk University)
14:00 - 14:35	Rethinking Learned Index and LSM-tree Integration	Jongmoo Choi, Dankook University
14:35 - 15:10	MemSOS: OS-Guided Selective Memory Mirroring	Jae Wook Lee, Seoul National University
15:10 - 15:45	DeLFS: A Decentralized Log-Structured File System for Manycores	Yongseok Son, Chung-Ang University
15:45 - 16:10	Break
	Session II HBF Special Session (Session Chair: Sungjin Lee, POSTECH)
16:10 - 16:30	Exploring High-Bandwidth Flash for Modern LLM Inference: Opportunities and Challenges	Jisung Park, POSTECH
16:30 - 16:50	Architectural Design of Low-latency and High-bandwidth 3D NAND with Split-Stacked Row Decoder for LLM Inference	Wonbo Shim, SEOULTECH
16:50 - 17:10	H3: Hybrid Architecture Using High Bandwidth Memory and High Bandwidth Flash for Cost-Efficient LLM Inference	Minho Ha, SK Hynix
17:10 - 17:40	Panel Discussion
17:40 - 18:00	Opening Ceremony SIGFAST Test-of-Time Paper Awards Ceremony*

Oct 23 (Fri): TBD
Time	Subject	Speaker
- 10:00	Breakfast (Provided by NVRAMOS)
	Session III (Session Chair: Sang-Hoon Kim, Ajou University)
10:00 - 10:35	BoostX™-NTI: Fast, Scalable and Flexible Storage Architecture with NVMe/TCP Initiator Acceleration	Jangwoo Kim, Seoul National University
10:35 - 11:10	Rethinking the Bottleneck of LFS: Hidden Cost of Uncontrolled Slack Space Recycling	Jin Yong Ha, Ewha Womans University
11:10 - 11:45	Trust-V: Toward Secure and Reliable Storage for Trusted Execution Environments	Jinsoo Jang, Chungnam National University.
11:45 - 13:00	Lunch (Provided by NVRAMOS)
	Session IV (Session chair: Sungyong Ahn, Pusan National Univ.)
13:00 - 13:35	TBD	FADU
13:35 - 14:10	TBD	Samsung Electronics
14:10 - 14:45	The Potential of Semantic-aware Tiering in Optimizing Agentic AI Workflows	SK Hynix
14:45 - 15:00	Break
	Session V (Session chair: Wookhee Kim, Konkuk Univ.)
15:00 - 15:35	SMOOTH: Hardware-Assisted Fine-Grained On-Chip Memory Management for Efficient On-Device LLM Inference	Daehoon Kim, Yonsei University
15:35 - 16:10	A Cost-Effective Near-Storage Processing Solution for Offline Inference of Long-Context LLMs	Jinho Lee, Seoul National University
16:10 - 16:45	Revisiting Pipeline Parallelism for LLM Serving	Jeongseob Ahn, Korea University
16:45 - 17:10	Closing Remark

Oct 24 (Sat): TBD
Time	Subject	Speaker
09:00 - 11:30	Panel Discussion: Future of the Storage Technology

*This award was established through the generous donation of the entire prize money received by Professor Youjip Won (KAIST) and Sang-Won Lee (SNU) upon winning the Gaheon Academic Award (가헌학술상) from KIISE in 2016 and 2022, respectively.