No disadvantage to two ports?

Ned Freed <NED@innosoft.com> Mon, 14 September 1992 05:23 UTC

Received: from NRI.NRI.Reston.Va.US by IETF.NRI.Reston.VA.US id aa10756; 14 Sep 92 1:23 EDT
Received: from ietf.NRI.Reston.Va.US by NRI.Reston.VA.US id aa03954; 14 Sep 92 1:27 EDT
Received: from ietf.nri.reston.va.us by IETF.NRI.Reston.VA.US id aa10748; 14 Sep 92 1:23 EDT
Received: from NRI.NRI.Reston.Va.US by IETF.NRI.Reston.VA.US id aa10723; 14 Sep 92 1:21 EDT
Received: from venera.isi.edu by NRI.Reston.VA.US id aa03895; 14 Sep 92 1:24 EDT
Received: from THOR.INNOSOFT.COM ([192.160.253.66]) by venera.isi.edu (5.65c/5.65+local-6) id <AA15907>; Sun, 13 Sep 1992 22:24:59 -0700
Received: from INNOSOFT.COM by INNOSOFT.COM (PMDF #1336 ) id <01GOR2WNMDM89FM7LA@INNOSOFT.COM>; Sun, 13 Sep 1992 22:24:27 PDT
Date: Sun, 13 Sep 1992 22:24:27 -0700
From: Ned Freed <NED@innosoft.com>
Subject: No disadvantage to two ports?
To: brnstnd@kramden.acf.nyu.edu
Cc: ietf@isi.edu
Message-Id: <01GOR2WNMN9E9FM7LA@INNOSOFT.COM>
X-Vms-To: IN%"brnstnd@kramden.acf.nyu.edu"
X-Vms-Cc: IN%"ietf@isi.edu"
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; CHARSET="US-ASCII"
Content-Transfer-Encoding: 7bit

Dan Bernstein writes:

> Actually, the reason that IDENT should be moved to a different port (and
> that fifteen or so people have said as much on the IETF list) is mainly
> one of practice. There *is* significant existing use of port 113, enough
> to render any incompatible protocol (such as IDENT) entirely useless.

The notion that IDENT is incompatible is far from established. Even if it is
incompatible recent postings indicate that the only incompatibilities are minor
and probably can be fixed.

> There is no *disadvantage* of moving IDENT. Thus it should be moved.

As has been pointed out before, there is a significant disadvantage to having
two protocols deployed on different ports that do the same thing. It makes it
necessary for clients to try both to be sure that all possible information is
obtained. The increase in effort is significant enough that it will probably
discourage many implementors.

I am opposed to changing ports since it is but one step further along the road
to two protocols on different ports. I don't much care which protocol we end up
with as long as we end up with only one.

Changing ports is simply a easy out. It defers the issue of which protocol to
use and support to implementors and the Internet community. Now I'm all for
"voting with our feet" when there are significantly different alternatives and
no clear criteria for making a choice. These protocols, however, do not even
come close to being significantly different. (The Ident MIB represents a good
example of a significantly different approach to this problem. It will be very
interesting to see whether the community eventually uses the MIB approach or
the RFC931/IDENT/TAP approach.)

There are things I don't expect the IETF/IESG/IAB to do for me. I don't expect
them to make choices for me when there are significant advantages to having
multiple different solutions. I do expect them to refuse to endorse multiple
nearly identical protocols just because it is easier to give in.

					Ned