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:20181221T160728Z
LOCATION:D163
DTSTART;TZID=America/Chicago:20181112T140000
DTEND;TZID=America/Chicago:20181112T142500
UID:submissions.supercomputing.org_SC18_sess142_ws_pdsw105@linklings.com
SUMMARY:Pufferbench: Evaluating and Optimizing Malleability of Distributed
  Storage
DESCRIPTION:Workshop\nI/O, Storage, Workshop Reg Pass\n\nPufferbench: Eval
 uating and Optimizing Malleability of Distributed Storage\n\nCheriere, Dor
 ier, Antoniu\n\nMalleability is the property of an application to be dynam
 ically rescaled at run time. It requires the possibility to dynamically ad
 d or remove resources to the infrastructure without interruption. Yet, man
 y Big Data applications cannot benefit from their inherent malleability, s
 ince their colocated distributed storage system is not malleable in practi
 ce. Commissioning or decommissioning storage nodes is generally assumed to
  be slow, as such operations have typically been designed for maintenance 
 only. New technologies, however, enable faster data transfers. Still, eval
 uating the performance of rescaling operations on a given platform is a ch
 allenge in itself: no tool currently exists for this purpose.\n\nWe introd
 uce Pufferbench, a benchmark for evaluating how fast one can scale up and 
 down a distributed storage system on a given infrastructure and, thereby, 
 how viably can one implement storage malleability on it. Besides, it can s
 erve to quickly prototype and evaluate mechanisms for malleability in exis
 ting distributed storage systems. We validate Pufferbench against theoreti
 cal lower bounds for commission and decommission: it can achieve performan
 ce within 16% of them. We use Pufferbench to evaluate in practice these op
 erations in HDFS: commission in HDFS could be accelerated by as much as 14
  times! Our results show that: (1) the lower bounds for commission and dec
 ommission times we previously established are sound and can be approached 
 in practice; (2) HDFS could handle these operations much more efficiently;
  most importantly, (3) malleability in distributed storage systems is viab
 le and should be further leveraged for Big Data applications.
URL:https://sc18.supercomputing.org/presentation/?id=ws_pdsw105&sess=sess1
 42
END:VEVENT
END:VCALENDAR