|
Folks,
Just
to add to this...
i) The
expedited flooding draft provides a fairly detailed comparison of signaling and
flooding for
time-constrained notification.
In fact, it provides several detailed network
examples to explain the requirements that sharing imposes,
and
then discusses some worst-case scenarios for both flooding and
signaling.
We
would be greatly interested in feedback on the discussion
therein.
It is
also worth emphasizing that the draft does not propose any specific way of
implementing flooding, and allows different proposals to achieve
it.
ii)
The FNP draft has now been restructured to separate the description of fault
timing in
optical transport networks from the details of the actual protocol. This
makes it easier
to see
exactly where in the recovery process the protocol comes into play, and
clearly
see
how the protocol operates.
Note
that FNP is implementation-agnostic, so it does not mandate that flooding be
realized
by any
particular protocol. Instead, it merely provides an abstract description of
the
elements needed for its operation, and allows for different realizations
of FNP.
iii)
Finally, the applicability draft is designed to provide a clear idea of how FNP
applies to optical
transport networks. So, it talks of the nodes, network, and fault models
for which FNP
currently works, and also discusses impact on network operation,
and on stability.
We
look forward to more feedback on the applicability draft.
Together the above drafts answer several questions raised by Roberto,
Alex, Neil, Zafar, Kireeti,
Jonathan (Sadler & Lang :-)), Adrian, etc. on
various aspects of the scheme and its operation,
and
allow for a new round of discussions.
-Vishal
|