[Olsr-users] OLSR Questions

[Markus Kittenberger]

Am 30.12.2009, 17:28 Uhr, schrieb Randy Buck <(spam-protected)>:

> Happy New Year!
same too!
> I have some questions that I'm hoping to get some help with.  Thanks.
> Randy
>
> 1) Why does the debugging output show INFINITE for ETX when it has  
> computed
> values for NLQ and LQ?
because INFINITE starts below 0.1 *G
>
> 2) When OLSR deletes a route and then adds a route, why is there  
> sometimes a
> delay of 2+ seconds between deleting and adding?  This could severely  
> impact
> TCP performance.
because olsrd is/was stupid in this case,..

did u find this in the newest version (tip of stable branch) as this  
should have been fixed 2 weeks ago,..
>
> 3) Why does OLSR sometimes change route when the mesh has no traffic?
why not?, olsrd does not care about traffic
>
> 4) Why does OLSR change routes when it gets only a minimal cost  
> improvement?
why not?, any received change is relevant, the only hysteresis in olsrd  
(only in new versions aswell) is done when sending packets

there used to be an imcoming hysteresis aswell, but i hope no parts of it  
are active any more,...
>
> 5) Why are routes either very good or very bad (cost of 68)?  Where does
> that high cost come from?  Is 60+ OLSR's definition of INFINITE?
maybe because your links are really either good or either bad, or because  
there might be be a bug as we "improved" the link-quality detection about  
2 months ago (to ensure it reacts faster on a died link)

please gibe us more informations which versions (hopefully only one *G)  
you are using, and try to measure the packetloss of the link yourself

a tcpdump capture file of all incoming olsr packets on this interface (at  
least from the specific neighbour) would be very helpful for analyze aswell
>
> 6) Why does link quality vary from very good to infinite for two nodes  
> that
> are adjacent?
i think this question can be answered together with 5) *G


-- 
lg Markus