You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@carbondata.apache.org by "Mohammad Shahid Khan (JIRA)" <ji...@apache.org> on 2016/07/28 11:14:20 UTC

[jira] [Created] (CARBONDATA-117) BlockLet distribution for optimum resource usage

Mohammad Shahid Khan created CARBONDATA-117:
-----------------------------------------------

             Summary: BlockLet distribution for optimum resource usage
                 Key: CARBONDATA-117
                 URL: https://issues.apache.org/jira/browse/CARBONDATA-117
             Project: CarbonData
          Issue Type: Improvement
          Components: Data Query
            Reporter: Mohammad Shahid Khan
            Assignee: Mohammad Shahid Khan


*Problem: *
In present scenario during querying data, the data scan can be done minimum at the block level.
Suppose if the system can execute 10 parallel task and number of blocks to be scanned is 3, then only 3 cores could be utilised remaining 7 will be un-utilized.
*Solution:*
To optimize the resouce usage the blocks can be further divided into multiple blocks.
A block consist of multiple blocklets, these blocklet could be distributed into multiple logical blocks by assigning the  startBlockletNumber and no of BlockLetsTobeScanned.

 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)