Mailing-List: contact mapreduce-issues-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: mapreduce-issues@hadoop.apache.org
Message-ID: <1248310748.318021266356248394.JavaMail.jira@brutus.apache.org>
Date: Tue, 16 Feb 2010 21:37:28 +0000 (UTC)
From: "Owen O'Malley (JIRA)" <jira@apache.org>
To: mapreduce-issues@hadoop.apache.org
Subject: [jira] Commented: (MAPREDUCE-326) The lowest level map-reduce APIs
 should be byte oriented
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/MAPREDUCE-326?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12834485#action_12834485 ] 

Owen O'Malley commented on MAPREDUCE-326:
-----------------------------------------

{quote}
One group should not block avenues of development because they are not their current priority when they might be a current priority for others.
{quote}

We haven't blocked anything. 

There are two proposals on the table:
* add an entirely new abstraction layer and API
* make a small incremental change to the current API

You insist on option 1. You haven't provided any arguments against the much simpler option 2.

Having implemented the data pipeline (sort and shuffle) multiple times, I don't think it is wise to expose the details of that implementation. They *will* change to address new issues as they arise. 

I think it is quite indicative of the relative merits that the people with the most experience in the data path for the sort and shuffle (Arun, Chris, and me) all have strong concerns about your proposal. 

We've spent a lot of time and effort helping to refine your proposal to get your goal accomplished and yet you continue to ignore our advice and experience. 

This is an Apache project where we need to reach consensus.

> The lowest level map-reduce APIs should be byte oriented
> --------------------------------------------------------
>
>                 Key: MAPREDUCE-326
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-326
>             Project: Hadoop Map/Reduce
>          Issue Type: Improvement
>            Reporter: eric baldeschwieler
>         Attachments: MAPREDUCE-326-api.patch, MAPREDUCE-326.pdf
>
>
> As discussed here:
> https://issues.apache.org/jira/browse/HADOOP-1986#action_12551237
> The templates, serializers and other complexities that allow map-reduce to use arbitrary types complicate the design and lead to lots of object creates and other overhead that a byte oriented design would not suffer.  I believe the lowest level implementation of hadoop map-reduce should have byte string oriented APIs (for keys and values).  This API would be more performant, simpler and more easily cross language.
> The existing API could be maintained as a thin layer on top of the leaner API.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.