[ 
https://issues.apache.org/jira/browse/SYNAPSE-408?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12619341#action_12619341
 ] 

Saliya Ekanayake commented on SYNAPSE-408:
------------------------------------------

It seems that the latest patch in the 
https://issues.apache.org/jira/browse/SYNAPSE-392 is not applied yet into the 
code base. Anyway, your point on CountingOutputStream and NullOutputStream is 
interesting and I'll try to give it a test soon. 

Thanks,
Saliya

> Improve Synapse Memory Footprint under HTTP 1.0
> -----------------------------------------------
>
>                 Key: SYNAPSE-408
>                 URL: https://issues.apache.org/jira/browse/SYNAPSE-408
>             Project: Synapse
>          Issue Type: Improvement
>            Reporter: Saliya Ekanayake
>            Priority: Minor
>
> Synapse serializes the SOAP envelope in order to calculate the content length 
> of the message under HTTP 1.0. This serialization is done inside the class 
> Axis2HttpRequest which is inside the transports module. The serialized bytes 
> are then stored in the message context for future use by the 
> streamMessageContents() method inside the same class.
> Thus, the entire content of the SOAP envelope is stored inside the memory 
> leading to a possible out of memory situation when the XML data is large. A 
> solution to this would be to write some of the data to a permanent storage 
> (like hard disk) based on a threshold value. The TemporaryData class inside 
> the core module is a good solution to this kind of work. It would, however, 
> incur a cyclic dependency if used inside the Axis2HttpRequest class. 
> If we can resolve this issue (probably by moving the TemporaryData class to a 
> utility module and then making the core too depend on that) it will of great 
> use in improving the memory footprint of Synapse.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to