[OLSR-users] Convincing olsr to select better routes

Amit Kucheria (spam-protected)
Thu Sep 30 23:05:44 CEST 2004


On Thu, 2004-09-30 at 13:26, aaron wrote:
> On Thu, 30 Sep 2004, Amit Kucheria wrote:
> 
> > On Thu, 2004-09-30 at 02:55, Michael Bauer wrote:
> >
> >> Now my question: is there any way we can convince the olsrd to select
> >> the routes via the (more hop, but better quality) directional route,
> >> instead of the omni route (few hops, bad quality)? We set willingness on
> >> the omnis to 1 and on directionals to 7, but that did not help.
> >
> > Certainly possible. But instead of forcing OLSR to behave differently
> > (from the RFC), how about adding an extension that gets RSSI from the
> > WLAN card. If this value is below a certain threshold, dont consider the
> > link in route calculations (give it less priority).
> 
> yes, but how do you deal with non-wlan interfaces?
> should it be like that?:
> 
> if (interface_type == wlan)
>    read_signal_strenght_stats();
>    if (stats < threshold)
>      ignore_this_interface();
> 
> ?
> 
> I guess this sounds like a nice --parameter startup option.

Obviously this only applied to wlan interfaces since I know of no way to
detect the 'quality' of an Ethernet link for example.

The option could be a config file and command-line option. I have been
thinking of ways in which settings can be dynamically modified e.g. when
the router detects no other 'links', it shouldn't discard the ones under
the threshold.

The big problem here is that WLAN cards donot seem to follow any
standards wrt providing signal strength readings. Last time I checked,
some provided it in dBm, others on a 1-100 scale, etc. That might have
changed though.

ciao,
Amit
-- 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://lists.olsr.org/pipermail/olsr-users/attachments/20040930/dd11fbc5/attachment.sig>


More information about the Olsr-users mailing list