[
https://issues.apache.org/jira/browse/HDFS-2058?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13047406#comment-13047406
]
Arun C Murthy commented on HDFS-2058:
-------------------------------------
{quote}
Protobuf has two nice features that thrift doesn't have yet: 1) when unknown
data is read, it is maintained in a map and then put back on the wire if the
same object is rewritten. 2) it has a decent plugin system that makes it easy
to modify the generated code – even with a plugin written in python or Java, in
theory. These could be implemented in Thrift, but again, I didn't want to take
the time.
{quote}
Similar reasons to choose PB for MRv2. Also, MRv2 has a pluggable layer, so in
theory, it could be anything. We choose PB for exact same reasons outlined by
Todd.
But, I'd encourage keeping serialization and in-memory data-structures separate
ala MR-279.
Mainly, we need to move ahead with something.
+1
> DataTransfer Protocol using protobufs
> -------------------------------------
>
> Key: HDFS-2058
> URL: https://issues.apache.org/jira/browse/HDFS-2058
> Project: Hadoop HDFS
> Issue Type: New Feature
> Affects Versions: 0.23.0
> Reporter: Todd Lipcon
> Assignee: Todd Lipcon
> Fix For: 0.23.0
>
> Attachments: HDFS-2058.patch, hdfs-2058.txt, hdfs-2058.txt,
> hdfs-2058.txt
>
>
> We've been talking about this for a long time... would be nice to use
> something like protobufs or Thrift for some of our wire protocols.
> I knocked together a prototype of DataTransferProtocol on top of proto bufs
> that seems to work.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
