BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:GraphChi and GraphChi-DB: large-scale graph computation on just a 
 PC - Aapo Kyrölä (Facebook)
DTSTART:20150205T150000Z
DTEND:20150205T160000Z
UID:TALK56958@talks.cam.ac.uk
CONTACT:Eiko Yoneki
DESCRIPTION:Current systems for graph computation require a distributed co
 mputing cluster to handle very large real-world problems\, such as analysi
 s on social networks or the web graph. While distributed computational res
 ources have become more accessible\, developing distributed graph algorith
 ms still remains challenging\, especially to non-experts.\n\nWe present Gr
 aphChi (OSDI¹12)\, a  disk-based system for computing efficiently on grap
 hs with  billions of edges. By using a well-known method to break large gr
 aphs into small parts\, and a novel Parallel Sliding Windows algorithm\,  
 GraphChi is able to execute several advanced data mining\, graph mining\, 
 and machine learning algorithms on very large graphs\, using just a single
  consumer-level computer. We show\, through experiments  and theoretical a
 nalysis\, that GraphChi performs well on both SSDs and rotational hard dri
 ves.\n\nWe build on the basis of Parallel Sliding Windows to propose a new
  data structure\, the Partitioned Adjacency Lists\, which we use to design
  an online graph database\, GraphChi-DB. We demonstrate that\, on a single
  PC\, GraphChi-DB can process over one hundred thousand graph updates per 
 second\, while simultaneously performing computation.  GraphChi-DB compare
 s favorable to existing graph databases\, particularly on data that is muc
 h larger than the available memory.\n\nWe will then look at our research i
 n hindsight and discuss what is GraphChi good for\, and what it is not opt
 imal for\, and also discuss competing systems that have been proposed sinc
 e. I will conclude with some remarks on open research problems.\n\nBio: Aa
 po Kyrola (Ph.D. Carnegie Mellon University\, 2014) is a computer scientis
 t\, software engineer and startup entrepreneur. The main contribution of h
 is Ph.D. Thesis is GraphChi (OSDI ¹12)\, a disk-based graph computation s
 ystem that can handle as large graphs as many distributed systems\, but on
  just a single PC or laptop. Mr. Kyrola also worked in the GraphLab projec
 t (UAI 2010\, VLDB 2012) and proposed a parallel algorithm for L1-regulari
 zed minimization problems (ICML 2011).\nIn CMU\, he was advised by profess
 ors Carlos Guestrin and Guy Blelloch.\n\nPrior to his Ph.D. Studies\, in 2
 000 Mr. Kyrola co-founded Habbo Hotel (habbo.com)\, a pioneering web-based
  virtual world that at its peak attracted over 15 million monthly users an
 d 50 million euros of revenue. During his Ph.D. studies\, he co-founded an
 other startup that developed a novel automatic activity tracker app Moves 
 for iOS and Android. Moves was chosen as one of the best apps by both Appl
 e and Google in 2013.\nFacebook acquired Moves in 2014\, and now Mr. Kyrol
 a works as a software engineer in Facebook's London office.\n
LOCATION:FW26\, Computer Laboratory\, William Gates Builiding
END:VEVENT
END:VCALENDAR