🚀 go-pugleaf

Philip,

"Connected Status" is an anachronistic hold-over from the old ARPAnet/
Milnet days, and was to have been done away with over a year ago (see RFC1174).
Unfortunately, the people who have so far paid the piper for NIC services,
the U.S. Department of Defense, didn't see it that way when they had to
re-compete the NIC services contract (they agreed in principle to RFC1174,
but the procurement specifications had already been keyed to the old way
of doing business).  For that reason, NSF agreed to grandfather in for
connected status sponsorship all the networks which had been included in
the inverse address tables since the brief adoption of the RFC1174 procedures
by the former NIC.  We also will be willing to sponsor networks  on a case-by-
case basis for connected status until the current procedures are abandoned
and RFC1174 is again implemented. To that end, NSF is preparing to issue a
competitive solicitation for Network Informatin Services, including Internet
Registration services in accordance with RFC1174 (=no more need for "Connected
Status").  It's almost as frustrating for us as is is for our colleagues
abroad, and it is downright confusing for network users.  Please be patient
with us for a little while longer.

In the meantime, if you have specific needs in this regard, please contact
NSF's Program Director for NSFNET, Dr. George Strawn <gstrawn@nsf.gov>.

					Thanks,

					Steve G.


  >All -- this is a general plea:
  >
  >	There is something that has been rankling me a while now,
  >and shows no sign of changing... that is the registration of
  >in-addr.arpa records for "non-connected" sites. Now the status
  >"non-connected" is a relative, or more correctly relativistic, term.
  >It depends on the perspective of the observer.
  >
  >	You might not consider random-french-network-X connected,
  >put that doesn't mean *I* can't send/receive packets to/from it.
  >And as a client of the root nameservers, I would like to be able
  >to resolve the name of those packets I might just happen to see
  >on my network, especially if it is someone trying to break into a
  >host.  Or a bogon source.
  >
  >	In any case, "connected" is an IP-level (and hence routing
  >policy) question.  It has nothing at all to do with name-to-
  >address or address-to-name translation (it's another layer and a
  >different ball of wax), and anyone who thinks otherwise should
  >(re)read RFC-814.
  >
  >	I sent mail to the NIC, they took weeks to answer, and sent
  >me a lame answer with no justification.
  >
  >	Two years ago, we broached this issue.  It was eventually,
  >and painfully, resolved on an individual basis.  It was a procedure
  >I don't want to go through again.
  >
  >	So, can we nix the chauvanist attitude that a network isn't
  >*really* connected unless it is connected to the US, and hence has
  >a "legitimate" reason to be resolvable??  IP has finally caught on
  >in Europe.  Really.  Even Japan.  As far away as Australia!!
  >
  >	If I can resolve a name to an address, why not the inverse?
  >
  >	Who came up with this policy, anyway?  I would like to ask
  >him/her to reconsider, or at least justify his/her reasoning...
  >
  >-Philip