Mailing-List: contact issues-help@flink.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@flink.apache.org
Date: Fri, 8 Dec 2017 10:58:00 +0000 (UTC)
From: "ASF GitHub Bot (JIRA)" <jira@apache.org>
To: issues@flink.apache.org
Message-ID: <JIRA.13123483.1512659333000.417161.1512730680074@Atlassian.JIRA>
In-Reply-To: <JIRA.13123483.1512659333000@Atlassian.JIRA>
References: <JIRA.13123483.1512659333000@Atlassian.JIRA> <JIRA.13123483.1512659333106@jira-lw-us.apache.org>
Subject: [jira] [Commented] (FLINK-8220) Implement set of network throughput
 benchmarks in Flink
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
archived-at: Fri, 08 Dec 2017 10:58:04 -0000


    [ https://issues.apache.org/jira/browse/FLINK-8220?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16283365#comment-16283365 ] 

ASF GitHub Bot commented on FLINK-8220:
---------------------------------------

Github user StephanEwen commented on the issue:

    https://github.com/apache/flink/pull/5134
  
    @greghogan That may or may not be possible, not 100% sure. Re-distributing it does certainly not work.
    
    To my current understanding, the Apache guidelines say we would need to make sure that downstream consumers of Flink (also the source) don't accidentally pull in a GPL license. For weak copyleft, where consuming it only via linking, it is okay, but for strong copyleft, where accidentally linking it results in copyleft, I think the hurdles for that are pretty high. There are some gray zone areas (like MPL) where the understanding is that "optional" dependency is okay because it requires an explicit action by a downstream user (re-adding the dependency) and the potential damage is not too high.
    
    I have seen a lot of cases first hand in the past months how many companies that want to use open source freak out as soon as they see an instance of GPL. We would make Flink's adoption easier if we did not have that.
    
    My first suggestion would be to consolidate this in a separate `flink-benchmarks` repository. That would also keep the core Flink repository "clear" of GPL. I could check with Apache legal for a final answer.


> Implement set of network throughput benchmarks in Flink
> -------------------------------------------------------
>
>                 Key: FLINK-8220
>                 URL: https://issues.apache.org/jira/browse/FLINK-8220
>             Project: Flink
>          Issue Type: New Feature
>          Components: Network
>            Reporter: Piotr Nowojski
>            Assignee: Piotr Nowojski
>
> Benchmarks should be defined and implemented in flink project and they will be executed in {{flink-benchmarks}} project.
> Configurable parameters: number of record writers and number of channels.


--
This message was sent by Atlassian JIRA
(v6.4.14#64029)