[OLSR-users] link quality / forcing hops

Andreas Tønnesen (spam-protected)
Tue Apr 27 13:15:24 CEST 2004


Do you want statistics for OLSR control traffic or all traffic? If you 
want link statistics
then the best way is to query the driver of the interface as I do in the 
preliminary
link-layer notification code. As mentioned earlier this will brobably be 
integrated into
OLSR link-sensing later on.

If you are looking for OLSR traffic only - this could relativley easy be 
achieved with
a plugin... Creating a command-line tool to retrieve this info would not 
be too hard
either.

I could implement this - but right now I'm quite busy writing my 
thesis(on olsrd)
and I have some exams in may - so you'd have to be patient ;)

You or somebody else interested could perhaps do it? It would only be a 
matter of
keeping track of the sequence number in the OLSR header and log the loss for
the last n packages... And create a IPC interface for a application to 
fetch this info.

But I can look into it one day I can find the time.

- Andreas

(spam-protected) wrote:

>Selon Andreas Tønnesen <(spam-protected)>:
>
>  
>
>>The easy(but dirty) way to solve this is to just block the MAC address of
>>57 at 51(and vica verca - tough that's not actually neccesary).
>>Do:
>>iptables -A INPUT -m mac --mac-source MAC_ADDRESS -j DROP
>>
>>As far as I can see there is no way to prevent this in OLSR without adding
>>link quality testing. That is if all OLSR messages get trough. If some 
>>messages
>>are lost then hysteresis could be used - but you could still experience some
>>bursty connection... So I think the MAC blocking soultion is the 
>>simplest one!
>>
>>    
>>
>
>OK.
>
>So I would be interested in :
>
>- having statistical count of drop and received packets,
>- having a simple unix command to dump some information from OLSR demon.
>Perhaps with some arguments. This could be easily be integrated in embeded
>httpd server under Linksys Wrt54g.
>
>I would like to see how many packets are lost from poor links and from good
>links. We will probably be able to make some assumptions after that.
>
>
>
> 
>  
>

-- 

-- 
Andreas Tønnesen((spam-protected))
UniK University Graduation Center
University of Oslo
http://www.olsr.org




More information about the Olsr-users mailing list