On-disk cache and TCP optimizations

[Juli Mallett]

Diego,

On Mon, Feb 13, 2012 at 09:15, Diego Woitasen <diego at woitasen.com.ar> wrote:
>  I'm writing this e-mail to announce that we (XTECH) were hired by
> the Credicoop Bank to write the code to support on-disk cache and TCP
> optimizations.  The bank uses a lot of open source software, they open
> source friendly, so they want to share this code with the community.
> They have almost 300 branches connected with a Linux based VPN, in the
> headquarters and in the end points. We want to add the WAN
> optimization to every VPN server.
>
>  The announce is because the bank wants to include this code in the
> trunk. In the next months, we are going to interact with the original
> authors to write the code following the coding guidelines, the current
> requirements of the project and everything else that is required for
> the code to be accepted.

This is very exciting stuff!  There's really only me to deal with
right now, and I'm pretty flexible :)  All kinds of models are mixed
in WANProxy, and there are new things I'm experimenting with all the
time, some of which end up in the tree, and most of which do not.
There are things that are half-finished or incomplete.  I would say
that as long as the code looks like it fits in with the rest of the
source tree there's no problem, but the rest of the source tree does
not all look the same.

As long as the license is the same (you, the bank, XCORP, etc., can
hold the copyright on individual new files, but the license terms must
be the same) I'm comfortable with having to do work to make it fit my
vision, or accept it as is if I am too lazy or too busy to do so.  I
know this will all be tremendously-useful to a wide variety of
WANProxy users, and thank you for your willingness to do the work,
thank the bank for their willingness to fund it, and all of you for
your intent to contribute it back.

>  We've been doing different tests with the current Wanproxy and the
> results are very promising (I have to translate them to English and
> I'll publish them soon) but some modifications are needed to be used
> in a scenario like the described above.

I look forward to your analyses and statistics, and the contribution
of your code!

Thanks,
Juli.