On Wed, 21 Sep 2005, Pashby, Ronald W CTR NSWCDD-B35 wrote:
2) Requiring all nodes implement Inverse Neighbor Discover with the
addidtion of the response holdoff timer.
The feature exists. But an all-nodes mandatory implementation
requirement is additional functionality, and I'm not sure
there's justification for that yet - but I admit that I did not
follow the discussion in the last meeting about this, so
I may be missing something. One approach would be
to publish INDbis spec, but not make it mandatory for
everyone.
As Jari said, I do not think adding mandatory-to-implement mechanisms
solves the problem. A solution already exists (to a degree), but the
vendors have voted with their feet (i.e., nobody actually wants the
feature).
In more general, I do not see sufficiently strong justification for
this approach. A typical operational approach in scenarios like these
is to collect the source IP addresses which send traffic at routers
(e.g., with logging access lists or netflow).
That approach can reliably identify any active node in the network --
mechanisms which rely on active probing (e.g., pings) will never be
able to find all the nodes anyway (e.g., because the hosts may filter
out pings in their local firewalls).
So, if you'd want to pursue the work on discovering the hosts in the
network, I'd put a lot more emphasis on describing why exactly this is
needed ("problem statement") and particularly why the existing
operational techniques are not sufficient.
--
Pekka Savola "You each name yourselves king, yet the
Netcore Oy kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
