Created 04-04-2018 11:16 AM
Hi,
Hope all doing well.
I'm looking for reason why data block size is 128 MB in Hadoop 2.x? What logic we have used to define the size should be 128 MB? Why we didn't define 100MB?
Created 04-11-2018 09:48 AM
The logic is quite simple: 128Mb is a multiple of "2" which means we can represent the number in binary like:
128Mb= 131072 Kb= 134217728 b = 1000000000000000000000000000 Binary
With this number we don't wast any bit when we stock data on memory
You can say that is a norme of storage of data in the computer science not just for big data
Created 04-05-2018 02:43 PM
Mainly it's because of performance reasons.
Have a read through this: https://community.hortonworks.com/questions/27567/write-performance-in-hdfs.html#answer-27633
Created 04-09-2018 08:04 AM
Thanks for reply.
But still i'm in doubt, Why it is not 126 MB or 132 MB?
Created 04-11-2018 09:48 AM
The logic is quite simple: 128Mb is a multiple of "2" which means we can represent the number in binary like:
128Mb= 131072 Kb= 134217728 b = 1000000000000000000000000000 Binary
With this number we don't wast any bit when we stock data on memory
You can say that is a norme of storage of data in the computer science not just for big data