For multiplexing only, I guess using SSH might be a bit overkill given
all the security aspects, though you can make those much lighter by
configuring the SSH protocol to not use enryption by using the "none"
cipher.  Though I don't know if there is any better protocol
specialized in multiplexing ...

I'll try to have another look at the port forwarding thing to see if i
can fix more issues.

On Wed, Oct 28, 2009 at 14:24, Lukas Nießen
<[email protected]> wrote:
> Am 27.10.2009 20:51 schrieb Guillaume Nodet:
>>
>> I've attached a new patch to SSHD-40 which addresses the stability
>> issues.   I've done a bit of testing and haven't found any problems
>> but those were not real load tests ...  Please have a look, i'm eager
>> for feedback on that one.
>
> Thanks for your work, great! From your comment on the patches-site, I
> followed that you integrated your solution into the main svn branch, so I
> checked out the svn version and used it instead of patching. I've done some
> testing, and it seems to work by far better than before. However, there
> still seems to be some (rather minor) issues. I did the same setup as
> described in my previous post, and many pages now load flawlessly, but
> others don't. These pages load most of their objects, but a few seem still
> to be missing. As far as I can see from the browser status bar, the missing
> objects are from some ad server or so. Examples for this behaviour include
> the pages
> www.focus.de
> www.yahoo.de
> www.bbc.co.uk/history/recent/troubles
> www.bahn.de
> (Sorry for being in German, but they belong to my test pages and it doesn't
> really matter here ;))
>
> Unfortunately, I've no clue where this behaviour may result from...
>
> Btw, as someone who should know the SSH protocol very good I ask you: Do you
> think it is a good idea to use the SSH protocol just for multiplexing (I
> don't need all the authentication and encryption stuff...) or do you happen
> to know a better suited protocol? If you're interested, I can of course
> describe in more detail what I'm willing to do.
>
> Again thank you very much, excellent work. Keep it up! And I'd be happy to
> contribute any stack traces whatsoever, but unfortunately, I'm not sure if
> I'm able to find the bug in the code... ;)
>
> Lukas
>



-- 
Cheers,
Guillaume Nodet
------------------------
Blog: http://gnodet.blogspot.com/
------------------------
Open Source SOA
http://fusesource.com

Reply via email to