[
https://issues.apache.org/jira/browse/FLINK-10718?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Flink Jira Bot updated FLINK-10718:
-----------------------------------
Labels: pull-request-available stale-major (was: pull-request-available)
> Use IO executor in RpcService for message serialization
> -------------------------------------------------------
>
> Key: FLINK-10718
> URL: https://issues.apache.org/jira/browse/FLINK-10718
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Task
> Affects Versions: 1.5.5, 1.6.2, 1.7.0
> Reporter: Till Rohrmann
> Priority: Major
> Labels: pull-request-available, stale-major
> Time Spent: 10m
> Remaining Estimate: 0h
>
> The {{AkkaInvocationHandler}} and once FLINK-10251 has been merged also the
> {{AkkaRpcActor}} serialize their remote RPC messages before sending them.
> This happens in the calling thread which can be a main thread of another
> {{RpcEndpoint}}. Depending on the de-/serialization time, this can be
> considered a blocking operation. Thus, I propose to introduce an IO executor
> which is being used for the serialization of the messages. This will make the
> {{RpcService}} abstraction more scalable.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
