[
https://issues.apache.org/jira/browse/HADOOP-7557?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13126173#comment-13126173
]
Sanjay Radia commented on HADOOP-7557:
--------------------------------------
I am interested in 2 and 3. I haven't stated otherwise. Indeed support for
non-java languages has been a very important motivation for a long time.
Hadoop RPC started to become pluggbale as part of the Avro thrust. Before I
started to work on this recently, Hadoop rpc already had mulitple RPC engines
(Writable, Avro, PB). The extra field (protocolKind) I added was to distinguish
between the multiple PRC Engines and catch bugs when there is a mismatch - this
should have been done when we added the multiple RPC engines.
We agree on the goal: "..select a single message serialization and
client/server protocol that's programming language-independent ..".
Given that the other layers are moving to PB or have already moved to PB
(HDFS's data transfer protocol and MR2 protocols) why not let the IPC header
be in PB?
Or are you proposing that all the upper protocol layers use a similar technique
to the one you used for the IpcHeader (Map<String,ByteBuffer>) and that we move
away from PB and Avro for Hadoop protocols?
> Make IPC header be extensible
> -------------------------------
>
> Key: HADOOP-7557
> URL: https://issues.apache.org/jira/browse/HADOOP-7557
> Project: Hadoop Common
> Issue Type: Improvement
> Reporter: Sanjay Radia
> Assignee: Sanjay Radia
> Attachments: HADOOP-7557.patch, IpcHeader.proto, ipcHeader1.patch,
> ipcHeader2.patch
>
>
--
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
