[Battlemesh] Out of sync seqnos after reboot [was: Battlemesh v5 tests]
jch at pps.jussieu.fr
Thu Mar 8 20:07:43 CET 2012
> We had a similar problem with the duplicate sequence number detection
> in OLSRd too for some time. We resolved it by counting how many
> "wrong" event we got without a single "right" one. After 8 events, we
> decide that source must have started with a new sequence number.
Yeah, we do that for hello seqnos in Babel, which are link-local.
Unfortunately, and unless I'm missing something, we cannot easily do
that for route seqnos, which are end-to-end, and can be delayed or
reordered by tens of seconds.
More information about the Battlemesh