[Olsr-users] Doubts about Willingness, MPR Selection and Routing.

Alex López (spam-protected)
Fri Feb 1 13:45:25 CET 2013


Hello,

After reading some documentation about OLSR and testing it with OLSRd in
OpenWRT I have some questions that I hope you can help me to answer.

First of all, about the MPR selection. I read the algorithm described in
RFC 3626 but when I try it in my lab, it doesn't work as expected (or maybe
I don't understand it well), for example:

- If set a 4 router network (R1, R2, R3, R4):

              /--R2--\
R1 --------/          \---------R4
             \---R3---/

Where *R1 and R4 cannot see each other directly and R2 and R3 are in the
same place. All routers have the same TX Power. Link Quality is set to 0.*

1.- The thing is that it doesn't matter which Willingness I set to R2 or R3
they are always selected as MPR (except when they have will=0). In theory,
if R2Will=6 and R3Will=1, shouldn’t be R2 the only MPR? Why to choose 2 MPR
instead of using only one and minimizing the flooding?

2.- Another thing I've found with this configuration is about the routing.
If I try to send a file from R1 to R4, all the data is always going through
R2 and is only going through R3 when R2Will=0. The routing shouldn’t select
the route taking into account the Willingness parameter as is said in the
RFC 3626?:

*          3.2  Several topology entries may be used to select a next hop
               R_next_addr for reaching the node R_dest_addr.  When h=1,
               ties should be broken such that nodes with highest
               willingness and MPR selectors are preferred as next hop.*


3.- In another different setup, where all the routers can see each other,
none of the routers is selected as an MPR, is this normal? Shouldn’t be at
least one MPR in a OLSR ad-hoc network?


Thanks in advance for the answers.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.olsr.org/pipermail/olsr-users/attachments/20130201/d43c1e67/attachment.html>


More information about the Olsr-users mailing list