BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:IC Colloquium : Big Data Analytics with Parallel Jobs
DTSTART:20130418T161500
DTEND:20130418T173000
DTSTAMP:20260408T085259Z
UID:0d98e3b21f9795bbf4ff9d454fd97908ed9b3864d2eabdded99e2d7f
CATEGORIES:Conferences - Seminars
DESCRIPTION:Ganesh Ananthanarayanan\, UC Berkeley\nIC faculty candidate\nA
 bstract\nExtensive data analysis has become the enabler for diagnostics an
 d decision making in many modern systems. These analyses have both competi
 tive as well as social benefits. To cope with the deluge in data that is g
 rowing faster than Moore’s law\, computation frameworks have resorted to
  massive parallelization of analytics jobs into many fine-grained tasks. T
 hese frameworks promised to provide efficient and fault-tolerant execution
  of these tasks. However\, meeting this promise in clusters spanning hundr
 eds of thousands of machines is challenging and a key departure from earli
 er work on parallel computing.\nA simple but key aspect of parallel jobs i
 s the all-or-nothing property: unless all tasks of a job are provided equa
 l improvement\, there is no speedup in the completion of the job. This tal
 k will demonstrate how the all-or-nothing property impacts replacement alg
 orithms in distributed caches for parallel jobs. Our coordinated caching s
 ystem\, PACMan\, makes global caching decisions and employs a provably opt
 imal cache replacement algorithm. A highlight of our evaluation using work
 loads from Facebook and Bing datacenters is that PACMan’s replacement al
 gorithm outperforms even Belady’s MIN (that uses an oracle) in speeding 
 up jobs. Along the way\, I will also describe how we broke the myth of dis
 k-locality’s importance in datacenter computing and solutions to mitigat
 e straggler tasks.Biography\nGanesh Ananthanarayanan is a PhD candidate in
  the University of California at Berkeley\, working with Prof. Ion Stoica 
 in the AMP Lab. His research interests are in systems and networking\, wit
 h a focus on cloud computing and large scale data analytics systems. Prior
  to joining Berkeley\, he worked for two years at Microsoft Research’s B
 angalore office.
LOCATION:BC 420 https://plan.epfl.ch/?room==BC%20420
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR