[
https://issues.apache.org/jira/browse/BAHIR-89?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15897099#comment-15897099
]
ASF GitHub Bot commented on BAHIR-89:
-------------------------------------
Github user fbeneventi commented on the issue:
https://github.com/apache/bahir/pull/37
@ckadner -- I wiil test the anntinutj PR in the next days, but I need to
repalce this in my current setup. I'm just using my own modified version of the
MQTT-connector that does quite the same of anntinutj but, however, it doesn't
support multi-topic input and so it wastes resources allocating multiple MQTT
receivers, one for each supplementary topic that has a different pattern (i.e.
cannot be inserted with wildcards).
The anntinutj PR is very important because:
- preserves resources in the Spark clusters
- gives the opportunity to filter topics, an thus to efficiently manage
payloads, inside the Spark programs.
> New API for subscribing from a list of MQTT topics and Return tuple of
> <Topic,Message> as output
> ------------------------------------------------------------------------------------------------
>
> Key: BAHIR-89
> URL: https://issues.apache.org/jira/browse/BAHIR-89
> Project: Bahir
> Issue Type: New Feature
> Components: Spark Streaming Connectors
> Affects Versions: Not Applicable
> Environment: Spark Streaming MQTT Connector
> Reporter: Anntinu Josy
> Labels: MQTT, SPARK
> Fix For: Not Applicable
>
> Original Estimate: 168h
> Remaining Estimate: 168h
>
> I am working in IoT Project. As part of MQTT-Kafka bridge program development
> I used Bahir. I feel that it will be a good feature to prove a new API to
> support a list of MQTT topic as input and output as a tuple of <Topic,
> Message>. This will be useful to reduce resource usage in case of multiple
> topic subscription. I had developed this feature a like to integrate
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
