[
https://issues.apache.org/jira/browse/MAPREDUCE-4049?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13415340#comment-13415340
]
Avner BenHanoch commented on MAPREDUCE-4049:
--------------------------------------------
Hi Arun,
With respect to your comments, please notice the following:
*regarding "MapOutputServlet to implement ShuffleProviderPlugin"*
1. I am doing that; however, please notice that at this moment,
MapOutputServlet will not be a plugin. Please let me know if you also want me
to make it a plugin.
2. making ShuffleProviderPlugin an interface (instead of abstract class as I
wrote it) will requires moving parts of its code to TaskTracker. (very few
things)
3. I have no problem to drop jobDone/taskDone from the interface. My core
plugin works without it. (it only served me as hints for cache optimization in
experimental variant of the plugin)
*We did some due-diligence on trunk:*
4. ShuffleConsumerPlugin - the interface will be implemented by the Shuffle
class. Main change is 'run' method instead of fetchOutputs & createKVIterator
methods. In Shuffle class, I moved CTOR work to an 'init' method.
5. ShuffleProviderPlugin is not needed in the trunk at all. The plugin will be
loaded as an AuxiliaryService. We already tried that.
*unit tests*
6. I'll add unit test(s) for both interfaces
7. this may delay the patch in few days since I never used JUnit (I mainly
program in C/C++). I am glad to learn it now. Hence, I may submit patch
without tests this week, and few days after that patch with tests.
8. I understand that I should mention my new test(s) in the mapred section of
smoke-tests file. Right?
Thank you for working with me on that,
Avner
> plugin for generic shuffle service
> ----------------------------------
>
> Key: MAPREDUCE-4049
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-4049
> Project: Hadoop Map/Reduce
> Issue Type: Improvement
> Components: performance, task, tasktracker
> Affects Versions: 1.0.3, 1.1.0, 2.0.0-alpha, 3.0.0
> Reporter: Avner BenHanoch
> Labels: merge, plugin, rdma, shuffle
> Attachments: HADOOP-1.0.2.patch, HADOOP-1.0.x.patch,
> HADOOP-1.1.patch, HADOOP-1.x.y-review-oriented.patch, Hadoop Shuffle Consumer
> Plugin TLD.rtf, Hadoop Shuffle Provider Plugin TLD.rtf, mapred-site.xml
>
>
> Support generic shuffle service as set of two plugins: ShuffleProvider &
> ShuffleConsumer.
> This will satisfy the following needs:
> # Better shuffle and merge performance. For example: we are working on
> shuffle plugin that performs shuffle over RDMA in fast networks (10gE, 40gE,
> or Infiniband) instead of using the current HTTP shuffle. Based on the fast
> RDMA shuffle, the plugin can also utilize a suitable merge approach during
> the intermediate merges. Hence, getting much better performance.
> # Satisfy MAPREDUCE-3060 - generic shuffle service for avoiding hidden
> dependency of NodeManager with a specific version of mapreduce shuffle
> (currently targeted to 0.24.0).
> References:
> # Hadoop Acceleration through Network Levitated Merging, by Prof. Weikuan Yu
> from Auburn University with others,
> [http://pasl.eng.auburn.edu/pubs/sc11-netlev.pdf]
> # I am attaching 2 documents with suggested Top Level Design for both plugins
> (currently, based on 1.0 branch)
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
