[Olsr-users] Two Questions: Dot Draw Plugin and ETX

[Aaron Kaplan]

Dave,


On Nov 27, 2007, at 9:19 AM, David Murray wrote:

> Hi,
>
> I have a couple of questions, I hope no one minds me asking two at  
> once.
>
> Firstly, I am playing around with the dot draw plugin at the moment.
> Does this plugin simply output the routing table in a format that can
> be drawn by the dot program? I am not complaining, It seems to work
> really well, but I think that if you are outputting the routing table
> from the perspective of one node then you are not getting the entire
> topology. It seems to me that you will only get the best path to each
> node in the network from the perspective of one node. Is this correct?

yes, an inherent property of link state protocols. you can say "alas"  
or "thank god!" depends on your perspective :)

>
> Another question I have is about the use of ETX (Expected Transmission
> Count). Since the work that introduced ETX
> (http://pdos.csail.mit.edu/papers/grid:mobicom03/paper.pdf) there have
> been improvements proposed such as ETT (Expected Transmission Time).
> This seems to improve ETX by adding the ability to measure bandwidth
> (described http://research.microsoft.com/mesh/papers/multiradio.pdf).
> Surely the ability to inculde bandwidth into the metric would be a  
> good
> addition. Is there any reason why ETX has not been adopted by OLSR? Do
                                   ^^^^^ you mean ETT here?
> people not want the added complexity or are there technical reasons  
> why
> it wouldn't work.
>

In fact, we are already discussing integrating ETT into olsrd.

care to help?

a.

(hehe: i might quote bernd now and say "send patches" )