> A discussion about how the use of the ill_t is changed would be
 > very beneficial.  At present there is an implication of the ill_t being
 > used for send/receive being the same as the information displayed
 > via ifconfig - your comments suggests that this is changed.

There is no such implication today.  For instance, if you configure an
IPMP group today in Nevada, packets may go out any functioning interface
in the group with a source data address taken from any interface in the
group.

 > Are you able to at least provide an itemised list of the low-level
 > architectural changes (without going into detail) so that we know
 > which areas have been considered and ticked off?

This sort of information will be part of the low-level write-up.

-- 
meem

Reply via email to