BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260202T201806Z
LOCATION:230
DTSTART;TZID=America/Chicago:20251117T110000
DTEND;TZID=America/Chicago:20251117T113000
UID:submissions.supercomputing.org_SC25_sess202_ws_pdsw109@linklings.com
SUMMARY:SlimIO: Lightweight I/O Path Design for Write Isolation in FDP-bac
 ked In-Memory Databases
DESCRIPTION:Sangyun Lee, Sungjin Byeon, Soon Hwang, and Jaewan Park (Sogan
 g University); Joo-Young Hwang, Junyoung Han, and Javier González (Samsung
 ); Awais Khan (Oak Ridge National Laboratory); and Youngjae Kim (Sogang Un
 iversity)\n\nIn-Memory Databases (IMDBs) are widely used with HPC applicat
 ions to manage transient data, often using snapshot-based persistence for 
 backups. Redis, a representative IMDB, employs both snapshot and Write-Ahe
 ad Log (WAL) mechanisms, storing data on persistent devices via the tradit
 ional kernel I/O path. This method incurs syscall overhead, I/O contention
  between processes, and SSD garbage collection (GC) delays. To address the
 se issues, we propose SlimIO, which adopts I/O passthru to minimize syscal
 l overhead and inter-process I/O interference. \nAdditionally, it leverage
 s Flexible Data Placement (FDP) SSDs as backup storage to avoid performanc
 e degradation from SSD GC. Experimental results show that SlimIO reduces s
 napshot time by up to 25%, increases query throughput by up to 30% during 
 non-snapshot periods, and lowers 99.9%-ile latency by up to 50%. Furthermo
 re, it achieves a write amplification factor (WAF) of 1.00, indicating no 
 redundant internal writes, thus extending SSD lifespan.\n\nTag: Data Analy
 tics, High Performance I/O, Storage, Archive, & File Systems, Storage\n\nR
 ecording: Livestreamed, Recorded\n\nRegistration Category: Technical Progr
 am Reg Pass, Workshop Reg Pass\n\nSession Chairs: Suren Byna (The Ohio Sta
 te University, Lawrence Berkeley National Laboratory (LBNL)); Anthony Koug
 kas (Illinois Institute of Technology, Argonne National Laboratory (ANL));
  Sarah Neuwirth (Johannes Gutenberg University Mainz); Ricardo Macedo (INE
 SC TEC, University of Minho); Jay Lofstead (Sandia National Laboratories, 
 University of New Mexico); and Dean Hildebrand (Google LLC)\n\n
END:VEVENT
END:VCALENDAR
