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:20260522T150123Z
LOCATION:C146
DTSTART;TZID=America/Chicago:20181114T140000
DTEND;TZID=America/Chicago:20181114T143000
UID:submissions.supercomputing.org_SC18_sess216_pap381@linklings.com
SUMMARY:Partial Redundancy in HPC Systems with Non-Uniform Node Reliabilit
 ies
DESCRIPTION:Zaeem Hussain, Taieb Znati, and Rami Melhem (University of Pit
 tsburgh)\n\nWe study the usefulness of partial redundancy in HPC message p
 assing systems where individual node failure distributions are not identic
 al. Prior research works on fault tolerance have generally assumed identic
 al failure distributions for the nodes of the system. In such settings, pa
 rtial replication has never been shown to outperform the two extremes (ful
 l and no-replication) for any significant range of node counts. We argue t
 hat partial redundancy may provide the best performance under the more rea
 listic assumption of non-identical node failure distributions. We provide 
 theoretical results on arranging nodes with different reliability values a
 mong replicas such that system reliability is maximized. Moreover, using s
 ystem reliability to compute MTTI (mean-time-to-interrupt) and expected co
 mpletion time of a partially replicated system, we numerically determine t
 he optimal partial replication degree. Our results indicate that partial r
 eplication can be a more efficient alternative to full replication at syst
 em scales where Checkpoint/Restart alone is not sufficient.\n\nTag: Perfor
 mance, Resiliency, Tools\n\nRegistration Category: Tech Program Reg Pass\n
 \nSession Chair: Michèle Weiland (EPCC, The University of Edinburgh; The U
 niversity of Edinburgh)\n\n
END:VEVENT
END:VCALENDAR