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:20260202T201809Z
LOCATION:275
DTSTART;TZID=America/Chicago:20251119T141500
DTEND;TZID=America/Chicago:20251119T143700
UID:submissions.supercomputing.org_SC25_sess283_pap943@linklings.com
SUMMARY:gParaKV: A GPGPU-Accelerated Key-Value Separation-Based KV Store w
 ith Optimized Compaction and Garbage Collection
DESCRIPTION:Hui Sun (Anhui University); Xiangxiang Jiang (Ahhui University
 ); Xiao Qin (Auburn University); Song Jiang (University of Texas, Arlingto
 n); and Enhui Wang (Anhui University)\n\nLSM tree-based key-value stores a
 re widely deployed in modern cloud storage systems thanks to high data sto
 rage efficiency and retrieval capabilities. The compaction in the LSM tree
 , however, results in severe performance bottlenecks, especially in large-
 sized value cases. While key-value separation methods mitigate the perform
 ance bottlenecks caused by compaction, the existing methods do not fully a
 ddress merge-sorting during compaction and expensive garbage collection (G
 C). We propose gParaKV, a GPGPU-empowered KV store with a KV separation me
 chanism, leveraging the GPGPU parallel technology to accelerate merge-sort
 ing in compaction and GC. gParaKV embraces a GPGPU bitmap structure, paral
 lel data marking, and a parallel GC mechanism. These critical components c
 urtail the overhead of merge-sorting and GC by virtue of parallel computin
 g. We compare it with state-of-the-art KV stores under various workloads. 
 The experimental results show that gParaKV can improve the write performan
 ce and GC efficiency compared to existing key-value separation-based KV st
 ores.\n\nTag: Data Analytics, Visualization & Storage\n\nRecording: Livest
 reamed, Recorded\n\nRegistration Category: Technical Program Reg Pass\n\nS
 ession Chair: Ana Kupresanian (Lawrence Berkeley National Laboratory (LBNL
 ))\n\n
END:VEVENT
END:VCALENDAR