[Olsr-users] OLSR in Android...a good idea?

[Charles Boyd]

Hi Robert!

First, I think your idea of flagging portable devices on an OLSR network is
interesting and would be worth considering further. I spent most of this
summer (through GSoC) working on a mesh networking application for Android
which is built on OLSR. The project is open source, you can browse the
repository online here:

http://wlan-lj.net/browser/trunk/meshapp

And the wiki page for the project (which needs an update) is here:

http://wlan-lj.net/wiki/Podrobnosti/MeshApp

I am still maintaining this project, but I have been busy with moving and
school for the last few weeks so there have not been any changes to the code
base recently. I will have time this weekend to work on a few critical
issues that will make the application (hopefully) workable in some basic
sense. If you have any thoughts or suggestions, I would be very glad to hear
them.

I have much of the equipment needed to set up a testbed here in Boston,
> but so few people interested in mesh wifi enough to give their time. What
> a shame.
>

This is fantastic! I am about an hour and a half away from Boston (depending
on how fast one drives...) out in Amherst, MA -- I study mathematics and
computer science here at UMass. I too have some mesh network equipment, and
have been seriously entertaining the idea of building a small test bed
around here.

Something interesting I have recently learned about is that there is already
some existing infrastructure for such a network to be built here, and there
are hoards of students that would be interested in such a project. See the
following links:

http://prisms.cs.umass.edu/dome/amherst-community-wireless

http://prisms.cs.umass.edu/dome/umassdieselnet

http://prisms.cs.umass.edu/dome/dieselnet-buses

So there is already some basis for work to be done here. Unfortunately,
these websites do not have particularly recent information -- so I will have
to find somebody who works on the project to get more information about what
the situation is currently like.

Anyway, I would be willing to drive down to Boston and meet up with you at
some point soon if you would like to discuss building a mesh testbed, mesh
networking on Android, tweaking OLSR for mobile usage...etc.

It would be quite interesting to have mesh networks in Boston and Amherst,
especially considering the relatively large student population.

Best,
-Charles


On Thu, Sep 16, 2010 at 1:32 PM, Robert Keyes <(spam-protected)> wrote:

>
>
> On Wed, 15 Sep 2010, Henning Rogge wrote:
>
> > On Wed, Sep 15, 2010 at 20:01, Robert Keyes <(spam-protected)> wrote:
> >> I am going to put OLSR on my G1 at some point.
> >>
> >> But I wonder if OLSR is really the most appropriate mesh protocol to be
> >> used on a portable device. Portability would seem to indicate frequent
> >> change of routing, and the updates required might swamp a network. I
> >> haven't tried this yet, and I imagine there is some network tuning which
> >> could help minimize this, but a portable node joining a large network of
> >> stationary nodes may not be so welcome.
> > There should be some optimizations for a combined "fixed/mobile" mesh.
> > For example the mesh should prefer nodes with "infinite" powersupply
> > for routes and flooding routing data.
> >
> > The best case would be if the wireless node is not needed for
> > routing/flooding at all. In this case it does not need to send TCs, so
> > it doesn't really congest the network that much.
>
> Perhaps nodes running on Android should have some sort of 'portable' flag
> they send to identify themselves, until the rest of the work is done in
> optimizing for portable nodes.
>
> Since the node's own idea of its 'predicted availability' might have a
> wide variety of estimations, represented by numbers that may be real
> rather than discrete, perhaps these is a simple method to handle based
> upon algorithm. I'll have to give this some more thought.
>
> The idea of power supply for transmission is an important one. Even nodes
> which are fixed may be battery powered, and when their batteries are low
> other routes should have preference. But such nodes would have a fixed
> location; so there should be a different attribute to indicate
> power-limited versus fixed/mobile location. A mobile node located in a car
> is one which has an unlimited power supply but is not in a fixed location,
> where as when it is parked, its power supply becomes limited and its
> location becomes fixed. A cellphone becomes stationary and of uniited
> power when it is plugged into its charger.
>
> What I have said above is not an argument or thesis but just my thinking
> aloud, so to speak, thinking while typing.
>
> I have much of the equipment needed to set up a testbed here in Boston,
> but so few people interested in mesh wifi enough to give their time. What
> a shame.
>
> --
> Olsr-users mailing list
> (spam-protected)
> http://lists.olsr.org/mailman/listinfo/olsr-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.olsr.org/pipermail/olsr-users/attachments/20100916/442daf81/attachment.html>