Wanproxy question

[Juli Mallett]

On Thu, Jun 20, 2013 at 1:08 AM, Mattia Peirano
<mattia.peirano at centeropenmiddleware.com> wrote:
> Hi!
>
> I write you because I'm trying Wanproxy. I think it is a good work and I'm
> trying to better understand the real architecture of the software in order
> to contribute.
>
> I tested it in a virtual environment and I noticed a behavior that I think
> could be improved. The machine running the proxyclient does not make flow
> control over the tcp connection with the host.
> Here the topology of the scenario:
>                                                           _____
> Host -------- WANProxy Client-------| WAN |---------WANProxy
> Server------Server
>
>
> If I'm not wrong, it seems that the proxy reads and processes all packets
> coming from the TCP buffer, allocating memory dynamically when it needs to
> store packets.
>
> In my case, when I start a greedy traffic source, the machine running the
> optimizer does not limit the source flow and ends without memory. Could be
> possible to limit the amount of memory dedicated to the "buffer"?
>
> What do you think about this? I'm very interested to know your opinion.

You are right that this is something that can happen with WANProxy;
unfortunately we do not have a mechanism for limiting either (1) the
amount of buffered data, or (2) the size of the datastore.

The first should be an easy fix, the second slightly less-so.  My
current focus in WANProxy is on making it multi-threaded for a number
of reasons, not least of which is for a user-space TCP/IP stack which
is being added so that WANProxy can be used as a transparent proxy.
Once that is finished, I have a few other high-priority items, but the
two aforementioned are very high on my list.

Thanks,
Juli.