BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:DB Seminar: Platforms and Applications for “Big and Fast” Data
  Analytics
DTSTART:20141107T123000
DTEND:20141107T133000
DTSTAMP:20260603T230930Z
UID:ee6f7d60657723d636ec7bf186ba8abea446a2432514dd8580f7f2e4
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. Yanlei Diao http://people.cs.umass.edu/~yanlei/\nRecent
 ly there has been a significant interest in building big data systems that
  can handle not only “big data” but also “fast data” for analytics
 . Our work is strongly motivated by recent real-world case studies that po
 int to the need for a general\, unified data processing framework to suppo
 rt analytical queries with different latency requirements. Towards this go
 al\, our project is designed to transform the popular MapReduce computatio
 n model\, originally proposed for batch processing\, into distributed (nea
 r) real-time processing.\nIn this talk\, I start by examining the widely u
 sed Hadoop system and presenting a thorough analysis to understand the cau
 ses of high latency in Hadoop. I then present a number of necessary archit
 ectural changes\, as well as new resource configuration and optimization t
 echniques to meet user-specified latency requirements while maximizing thr
 oughput. Experiments using typical workloads in click stream analysis and 
 twitter feed analysis show that our techniques reduce the latency from ten
 s or hundreds of seconds in Hadoop to sub-second in our system\, with 2x-7
 x increase in throughput. Our system also outperforms state-of-the-art dis
 tributed stream systems\, Twitter Storm and Spark Streaming\, by a wide ma
 rgin. Finally\, I will show some initial results and challenges of support
 ing big and fast data analytics in the emerging domain of genomics.
LOCATION:INM10 http://plan.epfl.ch/?zoom=20&recenter_y=5863818.98707&recen
 ter_x=730608.81878&layerNodes=fonds\,batiments\,labels\,information\,parki
 ngs_publics\,arrets_metro\,transports_publics&floor=0&q=INM10
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR