[WLANware] GSoC 2011 - Active and Passive Measurements

[Michael Bredel]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Mitar,

thanks for you answer and the additional information :-)

Of course we know that we have to apply individually - and we are
aware of the consequences :-) Thus, we are going to divide the tasks
into two projects, propose them, and see what will happen. I think the
potential workload is still high enough to keep two persons busy.

We both have some general knowledge regarding routing protocols at
least for wired networks. In addition I have some - maybe outdated -
knowledge about Ad Hoc routing protocols, since I worked on an Ad Hoc
network simulator for my diploma thesis. However, refreshing or
acquiring some knowledge regarding mesh routing seems not a big
challenge - at least in theory. And from a theoretic point of view
there is no difference between theory and practice :-)

The idea of using available bandwidth estimation (AvBw) techniques to
improve routing is quite interesting as well and, best to my
knowledge, there have been some contributions e.g. in the field of
overlay routing i.e. peer-to-peer systems. However, measuring
available bandwidth in a wireless network still faces some severe
challenges even on a per link basis. (Actually, it faces challenges on
a wired link as well, but it gets even messier on wireless links.)
Thus, an integration into routing protocols would be a second step.
But it is definitely worth looking at it.

Regarding our ported software I have to admit that we have not
published them yet. However, we would like to do that and if you can
offer me some hooks on a process how to do it, we will provide the
improvements within the next days. Actually, that was one of the main
ideas when we started to think about an application to the GSoC - to
get in touch with people working on related systems, so that we can
communicate and benefit from each other. We would like to push some of
our research results to public.

Regarding the review process of the proposals, can you provide some
information on that? As far as I understand from reading the mailing
list, the ideas proposed on your idea-page, are just ideas but, yet,
there must not be a mentor who is responsible for it? So, who are the
mentors and what kind of background, e.g. within the project, or
educational, do they have?

Regards,
Michael



On 03/29/2011 04:27 PM, Mitar wrote:
> Hi!
>
> 2011/3/29 Michael Bredel <michael.bredel at ikt.uni-hannover.de>:
>> a colleague and I are very interested in joining your project for
>> the GSoC this year.
>
> I hope you understand that you should apply individually. And that
> it is not necessary that both of you (or any) is accepted.
>
>> We adapted the Bootloader (U-Boot) of the platform and ported
>> some software such as Rude/Crude, Cfengine2, and SystemTap to run
>> under OpenWrt according to our needs.
>
> I hope you have pushed this improvements upstream?
>
>> In addition, we can provide a method for passive available
>> bandwidth estimation out of packet traces, that, today, only
>> works offline. However, it will be beneficial to integrate such a
>> method for online monitoring which can be done using the OpenIMP
>> framework.
>
> This would also be interesting for integration into routing
> protocols. ;-)
>
>> So, what do you people think about these proposals? Do you have
>> any comments, suggestions, further ideas? We would be very glad
>> to get some feedback. Furthermore, it would be great to get in
>> touch with the mentor of this project to further discuss our
>> ideas, the project schedule, possible deliverables, etc.
>
> I think they are valid ideas and proposals. I like them and would
> like to have them. I am just not sure I myself could mentor you
> this year. I have too many other things this year.
>
> How much experience you have with mesh networks? Routing protocols
> used in them? Problems they might have? Because at the end we
> would like from all this data measuring performance of our networks
> to know how to improve our networks and also routing in there. I am
> not asking this because I would think you must have knowledge or
> experience with this but that this might be a good thing also to
> research a bit.
>
> I think it would be useful for you to ask also on Battlemesh
> mailing list. We gather on those events exactly to test performance
> of our networks. There is also much data already available, some
> scripts, tests and so on. You could check them, analyze them in
> advance, so that you would understand maybe things better.
> Definitely get in touch with people on the list and ask them for
> suggestions or/and ideas what exactly in their opinion should be
> worth testing. Maybe also prepare together tools for next battle
> mesh. So that you will be able to test on real testbeds all this
> you would develop.
>
> http://battlemesh.org/
>
> (I have no idea where this data is hiding on the page. Ask on the
> mailing list or check its archive.)
>
>
> Mitar


- -- 
Dipl.-Ing. Michael Bredel, Emmy Noether Group
Performance of Communication Networks

Institute of Communications Technology (IKT)
http://www.ikt.uni-hannover.de/

Leibniz Universitaet Hannover
Faculty of Electrical Engineering and Information Technology
Tel: +49 511 762-17898  Fax: +49 511 762-3030
Postal address: Leibniz Universität Hannover, IKT, Appelstr. 9A
D-30167 Hannover, Germany

Encrypted eMail welcome! GPG/PGP-Key: 0xFE373595
85FA C9A5 777A 6380 4DBB 25D4 3A57 B701 FE37 3595
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk2S5+IACgkQOle3Af43NZW45ACgtiapHOBXJvAR5CwRWexUxoY2
gJsAn2PP5iapywJcKYVI1z1gFuZCz8cc
=PeMA
-----END PGP SIGNATURE-----