You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@kafka.apache.org by "Guozhang Wang (JIRA)" <ji...@apache.org> on 2017/10/10 00:14:00 UTC

[jira] [Created] (KAFKA-6034) Streams DSL to Processor Topology Translation Improvements

Guozhang Wang created KAFKA-6034:
------------------------------------

             Summary: Streams DSL to Processor Topology Translation Improvements
                 Key: KAFKA-6034
                 URL: https://issues.apache.org/jira/browse/KAFKA-6034
             Project: Kafka
          Issue Type: Improvement
          Components: streams
            Reporter: Guozhang Wang


Today our DSL to topology translation mechanism is operator-by-operator, and hence is sub-optimal in the resulted topology. Some known issues:

1. The repartition topics for aggregations and joins may be duplicated, i.e. containing exactly the same data as other topics.
2. state store's changelog topics are duplicates of other repartition / to / through topics.

We'd better have improve our DSL translation with a global optimization goal for the number of internal topics as well as state stores, i.e. "logical plan optimization".



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