[WLANware] GSoC 2010: Project IP/UDP encapsulation Kernel module

[ZioPRoTo (Saverio Proto)]

> I want to point you to two things:
> in OLSR there is already an experimental feature called "smart-gw" which will replace the old "nat-threshhold" system. This >essentially will become an automatic tunnel establishing protocol to a NAT gw.
[..CUT..]

I would like not to confuse students writing the applications.

Smart-gw is OLSR feature that makes possible to configure a IP-in-IP
tunnel to deliver traffic to a gateway. It is a specific thing within
OLSR. Both ends of the tunnel are supposed to be routers running OLSR

The Project "IP/UDP encapsulation Kernel module" is an indipendent
project from OLSR.

The kernel today makes possible to create IP-in-IP of IP-in-GRE
tunnels using the "ip" tool. But if you want IP/UDP is not possible to
configure the kernel to have such a tunnel. You need some VPN
software, like tinc or OpenVPN, maybe you have to use encription even
if not needed.

Now, if we write this IP/UDP encapsulation module, generic for the
Kernel, this can be used on any network with any routing protocol. I
guess also the smart-gw feature can use that module instead of the
ip-ip.

Basically once we have the module it is enough to work on:
src/linux/kernel_tunnel.c (olsr stable tree)

and use the UDP/IP instead of IP/IP tunnel.

Now you will ask: why ?

Now a key idea is that if you tunnel over UDP, instead of IP, you can
send all the traffic of the mesh network on a endpoint over the
Internet (e.g. fast server) and aggregate the bandwidth of many slow
ADSL there.

So instead of sticking people to a gateway inside the mesh, just make
all the traffic to Internet be tunneled to a fast server and then from
there all the traffic exits with the same IP address.
Users see more bandwidth (especially in upload) and they exploit for
the better an eventual route flap :)
This is the scenario, sorry I did not have the time to write all this
before, moreover, I did not want to confuse the students with this
complex mesh scenarios that are useless for the implementation of the
kernel module ...

For any other detail ask Marco Bonola, he is the guy from Ninux that
had this idea. He reads this list.

ciao :)

Saverio