RE: Heartbeats draft (fwd)

[Paul Koning <pkoning@xxxxxxxxx>]

Chinna,

I would tend to agree with Chris.  We have seen the sort of problems
he mentioned in the real world.  In addition, less likely to be a
problem but still significant is the issue of self-stabilization.  I
know that not all protocols include this as an issue, but I have
learned to treat it as a mandatory property.

If you want to argue against the proposed mechanism, you might find it 
easier to convince people if you put forth an alternate proposal (with 
details of algorithms and protocols) that achieves the same ends with
more efficiency.  You mention, for example, that the SA database out
of sync problem is easy to solve by other ways, but the approach you
refer to in passing is one that has been considered before and found
risky and/or unworkable.  If you believe something has been
overlooked, please fill in the details.

	paul