INTERIM_MEETING_REPORT_
Reported by Mike Ritter/Apple
Minutes of the IP over AppleTalk Working Group (APPLEIP)
January 15, 1992
PPP/AppleTalk:
New drafts are available in the Internet-Drafts Directory, they will
also be posted on AppleLink in the Networking Standards Folder. RFC
1171 & 1172 have been rewritten into one document, with additions for
authentication and dial-back. Brad Parker has been writing the
AppleTalk specifics and has changed the address negotiation so it has a
better chance of working. The specifications should be reviewed and
commented on by those interested in AppleTalk and PPP.
The following specifications are available in the Internet-Drafts
Directory on nic.ddn.mil or nnsc.nsf.net
draft-ietf-pppext-appletalk-00.txt
draft-ietf-pppext-authentication-02.txt
draft-ietf-pppext-lcp-02.txt
MacIP:
John Veizades and Tom Evans are negotiating over the final draft and
will have it posted by mid-February.
MIBs:
People expressed interest in the following MIBs:
o AppleTalk MIB+ transports and configuration
o Local Talk Repeater
o AT over PPP
o AURP
o ARAP
o MacIP
o AFP
o Printer Server
o DecNet over LT
o Mac System MIB - informational RFC
Anyone interested on working on any of these MIBs should send their
names and a proposed Charter to the Apple-IP mailing list. Peter
Caswell and Garth Conboy from Pacer expressed intense interest in
working on an AppleTalk Services MIB (Print, FileShare, etc.) Some
1
others expressed an interest in the ARAP Server MIB. Apple is also
interested in helping out on these, please contact Mike Ritter at
MWRitter@applelink.apple.com and he will try to get the correct parties
together.
Karen Frisa from CMU has been volunteered to work on the AppleTalk MIB+
(by Steve Waldbusser). There was general consensus that this MIB should
cover all of Inside AppleTalk (except AFP) and that it should try to
address the problem of router configuration.
Authentication for the AppleTalk MIB and routers was discussed. It was
agreed that any serious attempts should implement the SNMP security
protocols. Since these are still in the draft stages, vendors discussed
what they are using today. Most routers implement a trusted IP address.
For SNMP over AppleTalk it was noted that, due to dynamic node
addressing, a trusted node address was unworkable, but that a trusted
net number range was basically equivalent to a trusted IP host list.
There is a ``connectathon'' at Apple which will test inter-operation of
AURP, ARAP, and SNMP over AppleTalk. Tools to exercise the
implementations are available from Apple. (For SNMP test tools (both
over IP and AppleTalk) contact BLEE@applelink.apple.com.) Mike Ritter
of Apple promised to try to get the source of the SNMP test tools
released. InterConn offered demo versions of their management console
product that does SNMP over DDP (and runs on a Macintosh). People who
have SNMP Agents should talk to InterConn about incorporating their MIBs
into the product.
AURP:
Alan Oppenheimer reported that AURP router operational experience with
several Universities was being set up. For AURP test tools contact
OPPENHEIME1@applelink.apple.com. He also has copies of a document for
the routing protocol and update state diagram.
Alan has withdrawn the AURP draft from the IETF standards track. He is
working on an AURP document with APDA that will be an Apple standard and
recommended that an RFC should be written that describes how to run AURP
on the Internet. The APDA document will also be published as an
informational RFC.
The MIB description for AURP was sent out to the list - please send
comments to Alan and the list.
AppleTalk Directorate:
John Veizades promised to be an open and unbiased Working Group Chair
and professed an ability to keep things confidential if requested. He
asked two questions:
1. What would an AppleTalk Directorate be? and
2. What is the process and what concerns do people have?
2
Frank Slaughter volunteered his thoughts. He was disgruntled because he
thought the Group was working on an open standard (AURP) and it turned
out to be an Apple standard. He does not want to work on Apple
standards, but does want to work on open standards. He said everyone
was standardizing on TCP/IP for wide-area connectivity because the IETF
process really works. He said the difference between vendor and IETF
standards are that an IETF standard has an open forum available to work
on it, try it out, and later it's declared officially done by the IAB. A
vendor specific standard has the vendor product and the protocol done at
the same time. He thought that to make AppleTalk really usable for wide
area connectivity and large networks the IETF open standards process
must be adhered to.
Greg Minshall, from Novell said that in the IETF this Group is unique in
that it is very vendor specific. The only reason there is interest in
making these protocols open is because of the number of Macintoshes
there are in the market. He suggest that Macintosh to local routers as
opposed to Router to Router would be a logical place to split the
ownership of the protocol. He also reminded people that what is done
here relates to all Apple products. Don't get too idealistic about how
working groups work and the problems that do or do not follow. Greg
also recommended that the Group split into separate working groups and
have a specific Charter for each topic. There is an attempt to do
everything in this one Group which is wrong.
John Veizades pointed out that splitting working groups into other
existing directorates may cause problems because the Directors don't
know anything about AppleTalk.
Jonathan Wenocur seconded the idea of splitting up into separate working
groups and recommended that when an issue comes up it should be made
clear where the future control of a protocol lies - will it be an IETF
protocol or will it be an Apple standard?
Frank Slaughter wanted the Group to make this statement: ``Things that
are done by this Working Group are open standards and that is what we
are working towards.'' There was general consensus that this was
correct.
Alan Oppenheimer said that AURP 2.0 could be a standard protocol of the
IETF, Apple can put it under the IETF's control.
Frank Slaughter said that he does not feel the need to have Apple's
stamp of approval for future work.
John Veizades summarized that the Working Group has carved out a niche
that it is willing to work on open standards for AppleTalk solutions
(under IETF control) and doesn't require Apple's approval. Apple can be
a participant just like any other vendor. There was general consensus
that this was correct.
Greg Minshall put forth that the IAB doesn't want to take over Apple
protocols that Apple wants to keep proprietary.
3
Bob Morgan asked if there were several working groups should there be an
AppleTalk Directorate that stays around to guide all of the working
groups in this area. Is there really a need for one?
Jonathan Wenocour said that a Directorate is important because of
AppleTalk integration into IP areas (AURP, KIP, CAP, etc.) The IETF
would feel safer with more control over AppleTalk, especially with the
possibility of large AppleTalk internets tunneling through the Internet,
but there is the potential for conflicts with Apple.
John Veizades pointed to the work being done for AppleTalk over PPP. He
said that this Group can do important work without touching on what's
inside AppleTalk or other Apple proprietary protocols. He said the
Group is wildly enthusiastic about being able to do work in areas that
are not Apple proprietary protocols and everyone agreed.
John Veizades recommended that someone should write up a Working Group
Charter for the following areas if they were interested: AppleTalk
interior (and exterior) routing protocols that scale better than the
present one, network management and MIBs, AppleTalk over PPP, AppleTalk
over a variety of link layers, wide-area naming, and configuration
management and any other areas that people thought needed work on.
Greg Minshall posed the question: Does Apple have an obligation to cede
a protocol or tell the Working Group that they are working in the same
area? He answered it by saying that it was impossible to expect this.
What working groups do are open standards, what Apple does can be Apple
standards. In addition he recommended that the working groups have
Charters and limit their discussions to the Agenda.
The next meeting will be held during with the IETF Plenary in San Diego
(March 16-20, 1992).
Attendees
Debbie Alsop Alsop.D@applelink.apple.com
Philipe Boulanger Access.priv@applelink.apple.com
Craig Brenner Brenner2@applelink.apple.com
Rich Brown Richard.E.Brown@DARTMOUTH.EDU
Jerome Calvo transware@Applelink.Apple.Com
Peter Caswell pfc@pacer.soft.com
Edward Ching eching@wc.novell.com
Cyrus Chow cchow@ames.arc.nasa.gov
Garth Conboy gc@pacer.soft.com
Joy Cordell J_CORDELL@VMSNET.ENET.DEC.COM
Donald Eastlake dee@ranger.enet.dec.com
Richard Ford 72510.553@compuserve.com
Karen Frisa karen.frisa@andrew.cmu.edu
John Gawf gawf@compatible.com
Bob Jeckell robert_jeckell@nso.3com.com
Grisha Kotlyar grisha@farallon.com
Fidelia Kuang kuang@apple.com
Louise Laier laierl@applelink.apple.com
Greg Merrell merrell@greg.enet.dec.com
4
Greg Minshall minshall@wc.novell.com
Robert Morgan morgan@jessica.stanford.edu
Alan Oppenheimer oppenheimer1@applelink.apple.com
J. Bradford Parker brad@cayman.com
Christopher Ranch cranch@novell.com
Michael Ritter mwritter@applelink.apple.com
Randy Ryals D2604@applelink.apple.com
John Schafer john_schafer@um.cc.umich.edu
Mike Schumacher ms@telebit.com
Frank Slaughter fgs@shiva.com
Evan Solley solley@applelink.apple.com
David S.A. Stine dstine@cisco.com
Michael Swan neon.sw@applelink.apple.com
Steve Sweeney steves@farallon.com
Bob Van Andel BVA@applelink.apple.com
John Veizades veizades@apple.com
Jonathan Wenocur jhw@shiva.com
5