Slide 1

Slide 1 text

Separating Hadoop Myths from Reality Rob Anderson

Slide 2

Slide 2 text

1   The  Myths  &  Reali.es   Surrounding  Hadoop     Rob  Anderson   VP  Systems  Engineering  

Slide 3

Slide 3 text

2   Sales   SCM   CRM   Public   Web  Logs   Produc7on   Data   Sensor     Data   Click   Streams   Loca7on   Social   Media   Billing   Enterprise   Data  Hub   Hadoop  Changes  Analy.cs   “Simple  algorithms  and  lots  of  data  trump   complex  models  ”   Halevy,    Norvig,  and    Pereira,  Google   IEEE  Intelligent  Systems    

Slide 4

Slide 4 text

3  

Slide 5

Slide 5 text

4  

Slide 6

Slide 6 text

5   Data   Warehouse   Volume   Variety   Velocity  

Slide 7

Slide 7 text

6  

Slide 8

Slide 8 text

7   Big Data is hard to move…because it’s BIG

Slide 9

Slide 9 text

8   What  was  the  genius  of  Hadoop?   §  Fueling  an  industry  revolu7on   by  providing  infinite  capability   to  store  and  process  big  data   §  Expanding  analy7cs  across  data   types   §  Compelling  economics   –   20  to  100X  more  cost  effec7ve   than  alterna7ves  

Slide 10

Slide 10 text

9  

Slide 11

Slide 11 text

10   Random  Wri.ng  in  MapR   S1 S2 S3 S5 S4 S1, S2, S4 S1, S3 S1, S4, S5 S2, S4, S5 S3 Client   wri.ng   data   CLDB   Ask  for   64M  block   Create  cont.   Picks  master   and  2  replica  slaves   Write   next  chunk   to  S2   S2, S3, S5 aZach  

Slide 12

Slide 12 text

11  

Slide 13

Slide 13 text

12   MapR   Spout   TwiZer   TwiZer      API   TwiZerLogger   Storm         MapR   Op7onal   MapReduce   DFS  

Slide 14

Slide 14 text

13   hZp://www.flickr.com/photos/onemoreshotrog/8085462024/  

Slide 15

Slide 15 text

14   Hadoop  Distribu.ons  

Slide 16

Slide 16 text

Hadoop:  The  Disrup.ve  Technology     at  the  Core  of  Big  Data  

Slide 17

Slide 17 text

16  

Slide 18

Slide 18 text

17   The  Reality  is     Architecture  MaHers  

Slide 19

Slide 19 text

MapR  Data  System   Architecture  Comparison   HBase   JVM   HDFS   JVM   ext3/ext4   Disks   Other  Distribu7ons   Disks   MapR  M7  

Slide 20

Slide 20 text

Architecture  Results   Results  with  other   distribu.ons   Results  with   MapR  M7  

Slide 21

Slide 21 text

20  

Slide 22

Slide 22 text

Produc.on  Success  with  Hadoop  

Slide 23

Slide 23 text

22   2000+   Nodes   Fortune  100  Retailer  

Slide 24

Slide 24 text

23   1000+  Nodes   Fortune  100  Financial  Services  Company  

Slide 25

Slide 25 text

24  

Slide 26

Slide 26 text

25   Produc7on  Hadoop  in     Waste  Management  

Slide 27

Slide 27 text

26   Suntory  whiskey  

Slide 28

Slide 28 text

27  

Slide 29

Slide 29 text

28   Unique  Iden.ty   Ini.a.ve,  India    

Slide 30

Slide 30 text

No content

Slide 31

Slide 31 text

30     Thank  you   Big  Data  Spain!  

Slide 32

Slide 32 text

No content