[ 
https://issues.apache.org/jira/browse/TEZ-1228?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14044970#comment-14044970
 ] 

Siddharth Seth commented on TEZ-1228:
-------------------------------------

Looks good, reducing the amount of data sent over the wire would be really good.

Couple of questions - mainly future enhancements.
- Do we need a mechanism to skip keys. e.g. A key is read and doesn't match 
whatever criteria the processor requires - would be useful to be able to skip 
directly to the next key rather than having to read all values. Likely requires 
an index somewhere - so that could add unnecessary overhead.
- Do we want a version optimized for cases where the key and value length are 
fixed (Int for example), or is this not seen very often.

> Prototype IFile : Define a memory & merge optimized vertex-intermediate file 
> format for Tez
> -------------------------------------------------------------------------------------------
>
>                 Key: TEZ-1228
>                 URL: https://issues.apache.org/jira/browse/TEZ-1228
>             Project: Apache Tez
>          Issue Type: Improvement
>            Reporter: Rajesh Balamohan
>              Labels: perfomance
>         Attachments: TEZ-1228-IFile.pdf
>
>
> The current vertex-intermediate format used all across Tez is a flat file of 
> variable length k,v pairs. For a significant number of use-cases, in 
> particular the sorted output phase, a large number of consecutive  identical 
> keys are found within the same stream. The IFile format ends up writing each 
> key out fully into the stream to generate (K,V) pairs instead of ordering it 
> into a more efficient K, {V1, .. Vn} list.
> This duplication of key data needs larger buffers to hold in memory and 
> requires comparison between keys known to be identical while doing a merge 
> sort.
> This bug tracks the building of a prototype IFile format which is optimized 
> for lower uncompressed sizes within memory buffers and less compute intensive 
> to perform merge sorts during the reducer phase.



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Reply via email to