[
https://issues.apache.org/jira/browse/SAMOA-16?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14541707#comment-14541707
]
ASF GitHub Bot commented on SAMOA-16:
-------------------------------------
Github user gdfm commented on the pull request:
https://github.com/apache/incubator-samoa/pull/11#issuecomment-101618723
Indeed, I see what you mean. Given that the feedback loop in Flink is
faster, the number of attempts to split should increase.
This is expected, but the number of such attempts is upper bounded by the
ones tried on the Local engine, where there is no delay between request of the
split criterion and response by the local statistics.
We already have some flow control to regulate the rate of ingestion in
PrequentialEvaluation. I'll play a bit with it to see what happens.
When you put the 2 seconds delay in the Flink Processors, what happens (I
guess) is that the whole data streams through a very rough, sub-optimal version
of the tree. So it's very fast, but the precision drops considerably because of
the artificial limit on the number of split attempts.
> Add an adapter for Apache Flink-Streaming
> -----------------------------------------
>
> Key: SAMOA-16
> URL: https://issues.apache.org/jira/browse/SAMOA-16
> Project: SAMOA
> Issue Type: New Feature
> Reporter: Paris Carbone
> Assignee: Gianmarco De Francisci Morales
>
> Apache Flink-Streaming is a new system for distributed stream processing
> built for unique and flexible high level stream transformations. A Flink
> adapter for Samoa should be able to translate a Samoa Task topology into
> Flink streaming transformations. Some of the challenges are the compositional
> topology support, circle detection and their translation to Flink iterations.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
