Open-Mesh

As announced earlier this year the B.A.T.M.A.N. project participated in the GSoC 2011 amongst other organizations under the Freifunk umbrella. We were granted two slots which we filled with interesting project proposals. Both students have started working on their project a couple of weeks ago. We asked each of them to introduce themselves and speak about their project goals:

Hello everybody, I'm Antonio Quartulli from Brindisi, Italy (usually 
UTC+1). I am 24 years old and I'm currently attending the second year 
of the Master's Course in Computer Science at the University of Trento 
(Italy). I'm specializing in "Systems & Networks" (that's the name of 
my current specialization :-) ) and I should get awarded at the end of 
September.

I entered the Wifi Mesh Network world about one year ago when I 
started to study for the "Nomadic Communication" university course. 
Later I participated to the WirelessBattleMeshv3 in Bracciano (Rome - 
Italy) and there I had the possibility to meet various people which 
work on mesh networking and on several open-source routing protocols. 
Getting in touch with the community has been the first step which 
convinced me to deepen this field and B.A.T.M.A.N.-Adv was the
protocol which most attracted me. 

I've been around in the B.A.T.M.A.N.-Adv community for about one year 
and this gave me the possibility to understand the protocol and the 
details behind it (of course I still have to learn A LOT!!).

Several ideas have been presented in the last period which are meant
to bring improvements and new feature to the protocol. Most of those 
are only concepts and still need an implementation so I did a step 
forward and decided to apply for a student slot in the Freifunk 
organization.

This is my first GSoC experience and I am really happy to participate 
as student. My project concerns B.A.T.M.A.N.-Advanced, a mesh network
routing protocol, and the ARP request/response mechanism.

Basically, before starting the first IP communication, every host in 
the network needs to retrieve the destination Layer-2 address and this 
operation is executed by means of broadcast messages. It is easy to 
understand that in medium/large networks this could lead to high delay 
due to packet losses, so delaying the IP communication.

Here comes the idea of using a Distributed Hash Table to store Layer-2
addresses in the network: nodes will not need to use broadcast packets
anymore, instead they can use unicast packets to reach the node which 
is known to store the requested information in the DHT.

More details and up-to-date news are available on this page:
https://www.open-mesh.org/projects/batman-adv/wiki/DistributedArpTable

Stay tuned :-)

Hi folks,

I am Linus Luessing from Luebeck (ok, some 'L's too many, but not 
necessarily my fault ;) ), a city in Northern Germany at the Baltic 
Sea. I am 23 years old and currently a Bachelor's degree student in 
Computer Science - which actually did not involve any mesh networking
so far at all. However I've been fascinated of the Freifunk idea since 
I first heard of it.

Therefore I got in touch with B.A.T.M.A.N., especially the kernel 
module B.A.T.M.A.N. Advanced, about three years ago where we have 
been playing a little with this kind of new technology and 
implementations and trying to guess what kind of weird stuff this 
tricky wifi layer was doing. And it was also a very vibrant impression 
of being somehow involved in an Open Source Software Project for the 
first time. I was quite astonished about the direct, personal support 
and open way of sharing experiences and giving feedback.

During the last years I got more and more addicted to this and also 
tried to join several events, like the Chaos Communication Congress,
the Wireless Battle Mesh or the Chemnitzer Linuxtage. For a lot of 
new, conceptual ideas it was just more feasible to chat about it 
face-to-face. And there were a lot of ideas in this still kind of new, 
experimental field :).

It has been much fun so far and it is a great opportunity and an 
honour to be able to make some more of these ideas a reality during 
the Google Summer Of Code 2011 and to contribute to the Linux 
kernel, meshing, the community and Open Source Software in general.

During the GSoC 2011 I am working on some modularization of the core
routing protocol, splitting the OGM protocol as known of BATMAN IV
into an ELP (Echo Location Protocol, responsible for link, one hop 
tasks), OGM (will be responsible for the routing between nodes only, 
without the knowledge of separate links underneath) and a new MGO 
protocol (Message Guided towards Originator, a mechanism to be able
to recover from sudden path degradation quicker) in BATMAN V.

The protocol specs (work in progress!) can be found here:
https://www.open-mesh.org/projects/batman-adv/wiki/ELP
https://www.open-mesh.org/projects/batman-adv/wiki/OGM

Let's make the world a better place, bit by bit :).

Cheers, Linus

Happy routing,

The B.A.T.M.A.N. team