BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260422T143138Z
LOCATION:B312
DTSTART;TZID=America/New_York:20241118T103000
DTEND;TZID=America/New_York:20241118T110000
UID:submissions.supercomputing.org_SC24_sess811_ws_ciss101@linklings.com
SUMMARY:Can Current SDS Controllers Scale To Modern HPC Infrastructures?
DESCRIPTION:Mariana Miranda (INESC TEC & University of Minho); Yusuke Tani
 mura and Jason Haga (National Institute of Advanced Industrial Science and
  Technology (AIST), Japan); Amit Ruhela, Stephen Lien Harrell, and John Ca
 zes (Texas Advanced Computing Center & University of Texas at Austin); and
  Ricardo Macedo, José Pereira, and João Paulo (INESC TEC & University of M
 inho)\n\nModern supercomputers host numerous jobs that compete for shared 
 storage resources, causing I/O interference and performance degradation. S
 olutions based on software-defined storage (SDS) emerged to address this i
 ssue by coordinating the storage environment through the enforcement of Qo
 S policies. However, these often fail to consider the scale of modern HPC 
 infrastructures.\nIn this work, we explore the advantages and shortcomings
  of state-of-the-art SDS solutions and highlight the scale of current prod
 uction clusters and their rising trends. Furthermore, we conduct the first
  experimental study that sheds new insights into the performance and scala
 bility of flat and hierarchical SDS control plane designs.\nOur results, u
 sing the Frontera supercomputer, show that a flat design with a single con
 troller can scale up to 2,500 nodes with an average control cycle latency 
 of 41 ms, while hierarchical designs can handle up to 10,000 nodes with an
  average latency ranging between 69 and 103 ms.\n\nTag: I/O, Storage, Arch
 ive\n\nRegistration Category: Workshop Reg Pass\n\nSession Chairs: Glenn B
 rook (Cornelis Networks, University of Tennessee); Clayton Hughes (Sandia 
 National Laboratories); Nalini Kumar (Intel Corporation); Hatem Ltaief (Ki
 ng Abdullah University of Science and Technology (KAUST)); David Martin (L
 awrence Berkeley National Laboratory (LBNL), Energy Sciences Network (ESne
 t)); and Amit Ruhela (Texas Advanced Computing Center (TACC), University o
 f Texas)\n\n
END:VEVENT
END:VCALENDAR