BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Understanding and Improving the Efficiency of Failure Resilience f
 or Big Data Frameworks - Florin Dinu\, Rice University
DTSTART:20130415T101500Z
DTEND:20130415T111500Z
UID:TALK44043@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:Big data processing frameworks (MapReduce\, Hadoop\, Dryad) ar
 e hugely popular today. A strong selling point is their ability to provide
  failure resilience guarantees. They can run computations to completion de
 spite occasional failures in the system. However\, an overlooked point has
  been the efficiency of the failure resilience provided. The vision of thi
 s work is that big data frameworks should not only finish computations und
 er failures but minimize the impact of the failures on the computation tim
 e.\n\nThe first part of the talk presents the first in-depth analysis of t
 he efficiency of the failure resilience provided by the popular Hadoop fra
 mework at the level of a single job. The results show that compute node fa
 ilures can lead to variable and unpredictable job running times.\nThe caus
 es behind these results are detailed in the talk. The second part of the t
 alk focuses on providing failure resilience at the level of multi-job comp
 utations. It presents the design\, implementation and evaluation of RCMP\,
  a MapReduce system based on the fundamental insight that using replicatio
 n as the main failure resilience strategy oftentimes leads to significant 
 and unnecessary increases in computation running time. In contrast\, RCMP 
 is designed to use job re-computation as a first-order failure resilience 
 strategy. RCMP enables re-computations that perform the minimum amount of 
 work and also maximizes the efficiency of the re-computation work that sti
 ll needs to be performed.
LOCATION:Small Lecture Theatre\, Microsoft Research Ltd\, 21 Station Road\
 , Cambridge\, CB1 2FB
END:VEVENT
END:VCALENDAR