Here is my exchange with
@kdnuggets I missing this tweet. Of course can interview. Data size under management is 3 бо�€щоbyte. Is traditional Kazakhstan data measure
— Big Data Borat (@BigDataBorat) September 21, 2012
kdnuggets:
@BigDataBorat боpщ (Borscht) is a Big Ukrainian soup, so Borshobyte is for measuring Big Data streams, no? How many Borshobytes in Twitter?
BigDataBorat:
@kdnuggets
Twitter data big but no is боpщоbyte can better measure as боpщкаbyte. BigDataBorat Lab have policy for never laugh at small data
kdnuggets:
@BigDataBorat so what is the best way to deal with Big Data which is bigger than 3 боpщоbytes?
BigDataBorat:
- At BigDataBorat lab all rack is arrange in 3 dimensions cube with 3 stack on top each other. Best scaling of world wide.
- In all case is important for solution scale linearly. In extreme case solution need scale cubically.
- For best scale is important for bring compute close to data. But too close and data get restraining order against compute.
@BigDataBorat What tools do you use when you work with small data? And how do you Hadoop in Kazakhstan?
BigDataBorat:
- Work with Hadoop hard problem is find Java and UNIX skill. BigDataBoratLab solve with novel way.
- Outside BigDataBoratLab have box say "Free AbstractClassFactoryImpl inside." Anyone lift box trap in falling cage.
- From there Java expert sent to dungeon for work on Map/Reduce code.
- Small data like sand, if no careful can get every place. Best tool for work is microscope.
You can pet a sheep but you no can petabyte. #bigdata
TGIF ! Stay tuned until next Friday !