[ 
https://issues.apache.org/jira/browse/STORM-1565?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15341671#comment-15341671
 ] 

ASF GitHub Bot commented on STORM-1565:
---------------------------------------

Github user dan-blanchard commented on the issue:

    https://github.com/apache/storm/pull/1136
  
    I'm not on the Storm dev mailing list yet—I guess I should probably fix 
that—but we actually already have a Python implementation of msgpack 
serialization in [pystorm](https://github.com/pystorm/pystorm), the Python 
multi-lang implementation that powers 
[streamparse](https://github.com/Parsely/streamparse).
    
    I've been meaning to propose for a while that instead of having your own 
default Python multi-lang implementation in Storm that very few people use 
(because it's not very Pythonic or production-ready), you should instead point 
people to use [pystorm](https://github.com/pystorm/pystorm) at the very least. 
It provides all the functionality that the `storm.py` provides but with 
documentation, exception handling, logging, etc.  pystorm is intended to be 
used as a starting point for more full-fledged Python-Storm interop libraries 
(like streamparse), but if you want to replace the bare bones provdided by 
`storm.py` with an enhanced version of the same thing, it seems like pystorm is 
the way to go.


> Multi-Lang Performance Improvements
> -----------------------------------
>
>                 Key: STORM-1565
>                 URL: https://issues.apache.org/jira/browse/STORM-1565
>             Project: Apache Storm
>          Issue Type: Improvement
>          Components: storm-core
>            Reporter: Xin Wang
>            Assignee: Xin Wang
>
> 1. add _org.apache.storm.multilang.MessagePackSerializer_
> 2. change default "topology.multilang.serializer" to _MessagePackSerializer_
> According to http://msgpack.org/ : It's like JSON, but fast and small.
> {quote}
> MessagePack is an efficient binary serialization format. It lets you exchange 
> data among multiple languages like JSON. But it's faster and smaller. Small 
> integers are encoded into a single byte, and typical short strings require 
> only one extra byte in addition to the strings themselves.
> {quote}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to