Greg O'Shea writes: The mobile node MUST NOT send Mobility Header
messages of a particular type to a particular correspondent node more
than MAX_UPDATE_RATE times within a second.

So a MN with #n > 3 bindings at CN (e.g. using multiple HoA) must
prolong handoff over n/3 seconds just to transmit its BUs. Not great.

-----------------

Jari Arkko writes:

I agree that this is a problem. However, we have previously
decided to move functionality related to efficient handling of
multiple home addresses out of the draft. So I wonder if it
would be better to leave the rate limitation as is, and deal
with a more efficient treatment of multiple home addresses
in another document when we get to that point.

We could also try to either remove the rate limitation, or
fix it so that it behaves better. However, I fear that
the first option could create unexpected congestion when
someone happens to configure a large number of home addresses.
As engineers, the second option may sound attractive to us, but
maybe we should resist the temptation -- I'm pretty sure
we can't make the *whole* problem go away by just improving
the rate at which BUs can be sent. We still have to send
multiple packets, for instance. I'd rather not have the
publication of the document be pending on the optimizations
for mobile nodes that have a large number of home addresses.

One thing we could do perhaps is to change the value
of MAX_UPDATE_RATE to a slightly higher value if this
presents a practical problem for planned deployments.

What do you think?

-----------------

Greg O'Shea writes: I'm not too worried by this one - it's just a bit
clumsy. I keep finding these things while working on an implementation
so I figure I might as well notify them to the list.

-----------------

-----------------

-----------------

-----------------

-----------------