multiple consumer of intermediate data set

[yoon]

Hi All,

 

figure1
https://ci.apache.org/projects/flink/flink-docs-release-1.2/fig/job_and_execution_graph.svg

 

as we can see in figure1, JobVertex(B) has two consumer( JobVertex(C) and JobVertex(D) )

and accordingly Intermediate Data Set of JobVertex(B) has two consumer( JobVertex(C) and JobVertex(D) )
but in case of JobVertex(A), though it has two consumer( JobVertex(B) and JobVertex(D) ) same as JobVertex(B)

it has two separate intermediates data set and each intermediate data set has one consumer.
i couldn't understand why... for me it looks same case but why one has one Intermediate Data Set and another has two?
could anyone explain what is difference between JobVertex(A) and JobVertex(B)? 

[Aljoscha Krettek]

For context, this is from the Flink doc: https://ci.apache.org/projects/flink/flink-docs-release-1.3/internals/job_scheduling.html

I think Ufuk (cc-ed) could know more about this>

On Mon, Mar 13, 2017, at 05:42, 윤형덕 wrote:

Hi All,

 

figure1

https://ci.apache.org/projects/flink/flink-docs-release-1.2/fig/job_and_execution_graph.svg

 

as we can see in figure1, JobVertex(B) has two consumer( JobVertex(C) and JobVertex(D) )

and accordingly Intermediate Data Set of JobVertex(B) has two consumer( JobVertex(C) and JobVertex(D) )

but in case of JobVertex(A), though it has two consumer( JobVertex(B) and JobVertex(D) ) same as JobVertex(B)

it has two separate intermediates data set and each intermediate data set has one consumer.

i couldn't understand why... for me it looks same case but why one has one Intermediate Data Set and another has two?

could anyone explain what is difference between JobVertex(A) and JobVertex(B)?