[Olsr-users] getting second best route from dikjstra's algorithm

[Vigneswaran R]

On 06/16/2013 07:37 PM, Juliusz Chroboczek wrote:
>> Is there any way (API) to get the second best route to a destination
>> from dikjtra's algorithm (when multiple routes are available between
>> source and the destination)?
> There's been quite a lot of work done on multipath routing in
> link-state protocols.  As Henning notes, the tricky bit is ensuring
> that the two routes are actually disjoint.
>
> I don't have my bibliography handy, but the following paper should
> contain some good references:
>
>    http://gyroweb.inria.fr/~viennot/publis/articles/sirocco10multipath.pdf

Ok.

>> The idea is to write a plugin to get the second (third etc.,) best
>> routes (for destinations) and keep them in separate routing tables.
> There's an experimental (unpublished) branch of babeld that does
> something very similar, except that we use the *source* address of
> packets to choose the route.  This puts the responsibility for
> splitting traffic in the application, which can choose a given path by
> picking a source address.  (We should be able to publish the code next
> week, you'll know everything if you follow babel-users.)

I didn't know about this branch. I will check this out.

> Since we're manipulating routing rules in order to achieve that, it
> should be possible to use different ways to choose the path.  What
> exactly do you have in mind?

As of now I don't have anything specific regarding implementation. We 
are looking into the suggestions made on this thread. Thanks for your 
inputs.


Regards,
Vignesh