Hi all,
Picking up on this comment from Thomas Clausen:
"But, if A is close to GW1, and B is close to GW2, but A and B being far
far apart (according to the metric used) within the MANET, then it would be
preferential for the route from A to B to go via GW1-“the-Internet”-GW2? "
I've drawn a little diagram to help visualise:
A-------------X----------Y--------Z-------B
| |
| (191.0.2.0/24) |
| |
| |
GW1 GW2
\ /
--------------Internet-------------------
I've worked through the details. That "preferential" route doesn't get
discovered and there are no sequence number issues, because only B is
allowed to send an RREP for itself, a GW is not allowed (since TargAddr is
B, and its inside the configured AODVv2 network, the GW doesn't respond).
Some extra details could be added to section 9 of the draft to clarify...
However,
*The problem actually occurs when A sends a RREQ for something on the
internet, since both GW1 and GW2 may respond. The one giving the highest
seqnum will be preferred with the current logic. This is equivalent to two
routers being responsible for one Router Client...which we disallowed.*
Any ideas before I send the big analysis to MANET without a solution...?
Kind regards,
Vicky.
