[
https://issues.apache.org/jira/browse/SSHD-164?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Pierre updated SSHD-164:
------------------------
Description:
InvertedShellWrapper uses by default a buffer of 512 bytes to copy the
input/output/error streams from the shell to the socket. For some applications,
this is way too low, as for example, copying 1MB from the process output to the
socket is going to require in the 2048 loop iterations in the IO pump thread.
For example, in our tests, switching the buffer size from 512 bytes to
32kBytes, on the transmission of 830Mb leads to an improvement from 1m30s for
the 512b buffer to 1m15s for the 32k buffer, saving of 1/6th of the overall
transmission time.
was:
InvertedShellWrapper uses by default a buffer of 512 bytes to copy the
input/output/error streams from the shell to the socket. For some applications,
this is way too low, as for example, copying 1MB from the process output to the
socket is going to require 2048 loop iterations in the IO pump thread, and
introduce a 2s of latency because of the 1ms sleep at the end of each loop
iteration.
Instead the buffer size should be configurable by the command factory.
> Allow the buffer size for the IO pumps to be configurable in
> InvertedShellWrapper
> ---------------------------------------------------------------------------------
>
> Key: SSHD-164
> URL: https://issues.apache.org/jira/browse/SSHD-164
> Project: MINA SSHD
> Issue Type: Improvement
> Affects Versions: 0.6.0
> Reporter: Pierre
> Priority: Minor
> Attachments: SSHD-164.patch
>
>
> InvertedShellWrapper uses by default a buffer of 512 bytes to copy the
> input/output/error streams from the shell to the socket. For some
> applications, this is way too low, as for example, copying 1MB from the
> process output to the socket is going to require in the 2048 loop iterations
> in the IO pump thread.
> For example, in our tests, switching the buffer size from 512 bytes to
> 32kBytes, on the transmission of 830Mb leads to an improvement from 1m30s for
> the 512b buffer to 1m15s for the 32k buffer, saving of 1/6th of the overall
> transmission time.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
