As a hadoop engineer or big data engineer or hadoop learner how can you explain what is hadoop to a kid?
This presentation makes the case for MapReduce (i.e. distributed computing) easy for anybody to understand:
MapReduce in Simple Terms
Hadoop is an implementation of MapReduce.
If they are still with you at this point, and they know that files are stored on a hard disk on a computer, you can ask what will happen to the files when you use 10 or 20 computers to process them? Should we copy all the files to all the computers locally? what if the files are too big for one single computer's hard disk storage? then we need to split the files up, and run a database to keep track of where each piece of each file is. that's the NameNode and HDFS.
so the more complete picture of Hadoop is MapReduce + HDFS.
Which brings us to MapR: The original version of HDFS is very slow and has a single point of failure in the NameNode that can kill a cluster just from that one server breaking down for any reason. So MapR founder M.C. Srivas decided to code a super fast, super robust version of HDFS that we call MapR-FS.
Let me know how much your 9 y.o. could get out of this very short story. :-)
Thank you for providing such a great information Mathieu Dumoulin
I also found below image related to easy way to understand map-reduce
Retrieving data ...