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:20210402T160210Z
LOCATION:Track 7
DTSTART;TZID=America/New_York:20201119T160000
DTEND;TZID=America/New_York:20201119T163000
UID:submissions.supercomputing.org_SC20_sess294_drs108@linklings.com
SUMMARY:A Dynamic Multi-Tiered Storage System for Extreme-Scale Computing
DESCRIPTION:Doctoral Showcase\n\nA Dynamic Multi-Tiered Storage System for
  Extreme-Scale Computing\n\nDevarajan, Sun\n\nTraditional scientific disco
 very was driven through the compute power of a computer system. Hence, I/O
  was treated as sparse tasks to perform occasional checkpoints for the app
 lication. This approach led to a growing gap between the compute power and
  storage capabilities. In the era of data explosion, however, where data a
 nalysis is essential for scientific discoveries, the slow storage system h
 as led to the research conundrum known as I/O bottleneck. Additionally, th
 e explosion of data has led to proliferation of application as well as sto
 rage technologies. This has created a complex matching problem between div
 erse application requirements and storage technology features. In this pro
 posal, we introduce Jal, a dynamic, re-configurable and heterogeneous-awar
 e storage system. Jal utilizes a layered approach that includes appli
 cation model, data model and storage model. The application model uses a s
 ource-code based profiler which identifies the cause of the I/O behavior o
 f applications. The data model translates various applications' I/O requir
 ements to an underlying storage configuration to extract maximum performan
 ce from each application. Finally, the storage model builds a heterogeneou
 s-aware storage system which can be dynamically re-configured to different
  storage configurations. Our evaluations have shown that these models can 
 accelerate I/O for the application while transparently and efficiently usi
 ng the diverse storage systems.\n\nRegistration Category: Tech Program Reg
  Pass
END:VEVENT
END:VCALENDAR