[Olsr-users] using dijkstra for route calculation

[Adrian Schwartz]

Incremented dijkstra sounds  intresting.
what about the performance of the old dijkstra?
is there a way to do performance evaluation of olsr in a single computer?
I would like to emulate 100 nodes and to check it by myself.

Another question,
In case of a link state change, how is the change flooded through the
network?

On Mon, Sep 20, 2010 at 4:17 PM, L. Aaron Kaplan <(spam-protected)> wrote:

>
>  On Sep 20, 2010, at 3:40 PM, Adrian Schwartz wrote:
>
>  I was wondering,
> how bad is it using dijkstra for full network routing calculation?
> lets assume that I have 100 nodes and I need to recalculate the path every
> second?
>
>
> By now deadly efficient!
> This was a big part of the OLSR-NG optimization work. Mainly Hannes
> optimized Dijkstra until it was so efficient that OLSRd only took a few
> percent CPU load on a small 200MHz embedded Linksys device which was
> calculating the shortest path in a up to 1000 node network.
>
> Right now there is a google summer of code project (we did not test it's
> results yet!) to have an incremental Dijkstra calculation (i.e.: only
> re-calculate the regions where things changed a lot)
>
> So, don't worry about CPU load right now on OLSR. We scale!
>
>
> a.
>
>   On Mon, Sep 20, 2010 at 1:58 PM, L. Aaron Kaplan <(spam-protected)>wrote:
>
>>
>> On Sep 20, 2010, at 1:18 PM, Adrian Schwartz wrote:
>>
>> > So how does the routing algorithm works?
>> > shortest path?
>> >
>> yes
>>
>>
> --
> 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/20100920/a8fc7666/attachment.html>