[
https://issues.apache.org/jira/browse/MAPREDUCE-4049?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Avner BenHanoch updated MAPREDUCE-4049:
---------------------------------------
Attachment: Hadoop Shuffle Plugin Design.rtf
Attached is the updated design documents.
For your convenience, I am writing below its essence:
* Support _ShuffleConsumerPlugin_ from 3rd parties which will be able to
retrieve maps from either the built-in _ShuffleHandler_ or from a 3rd party
Shuffle Provider Service (that can be loaded as an _AuxiliaryService_ at run
time)
* Update Hadoop code, so the current _Shuffle_ class will be the first (and
the default) _ShuffleConsumerPlugin_.
* Allow users to configure their desired _ShuffleConsumerPlugin_ in
_mapred.xml_ files.
* Change _ReduceTask_ to *dynamically load* the user configured
_ShuffleConsumerPlugin_ and use it for its shuffle needs.
* The interface of _ShuffleConsumerPlugin_ will be inspired by the interface
of the existing *Shuffle* class. *Hence, there will be minimal changes in
current code.*
{code:title=ShuffleConsumerPlugin.java|borderStyle=solid}
/**
* ShuffleConsumerPlugin for serving Reducers. It may shuffle MOF files from
* either the built-in ShuffleHandler or from a 3rd party AuxiliaryService.
*/
public abstract class ShuffleConsumerPlugin<K, V> {
public abstract void init(ShuffleContext<K, V> context);
public abstract RawKeyValueIterator run() throws IOException,
InterruptedException;
public void close(){}
/**
* Factory method for getting a ShuffleConsumerPlugin from the given class
* object and configuring it. If clazz is null, this method will return an
* instance of 'Shuffle' class since it is the default ShuffleConsumerPlugin
*
* @param clazz class of the requested ShuffleConsumerPlugin
* @param conf configure the plugin with this
* @return an instance of ShuffleConsumerPlugin
*/
public static ShuffleConsumerPlugin getShuffleConsumerPlugin(
Class<? extends ShuffleConsumerPlugin> clazz,
JobConf conf) throws ClassNotFoundException, IOException {
// ...
}
}
{code}
> 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.x.y.patch, Hadoop Shuffle Plugin Design.rtf,
> mapreduce-4049.patch, mapreduce-4049.patch
>
>
> 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
For more information on JIRA, see: http://www.atlassian.com/software/jira
