Date: Tue, 05 Jan 93 08:56:57 GMT
From: Jill.Foster@newcastle.ac.uk
Subject: IETF Trip Report
To: nir@cc.mcgill.ca
Cc: nir@mailbase.ac.uk
This report is about 1000 lines long - but is structured to allow you
to skip over bits not of interest to you. I started to edit out bits
that weren't "NIR" specific - but found bits in discussions of NISI
etc that were of potential interest to some of you. You'll probably
want to skip most of the Plenary stuff.
Best wishes for a constructive year in NIR!
-- Jill
IETF - Washington: Nov 16-20, 1992
==================================
Trip Report:
Jill Foster - Newcastle University, UK
Chairman: RARE Information Services and User Support Working Group
Introduction
============
The 25th IETF (Internet Engineering Task Force) Meeting took place in
Washington DC from 16th-20th November and was hosted by SPRINT.
Silver coloured "25th IETF" buttons were on sale to mark the occasion.
The attendance was just under 600, which was down slightly on the
Boston IETF.
My main reasons for attending (thanks to funding from RARE) were to:
o represent RARE Information Services and User Support Working
Group (which I chair)
[RARE is the Association of European Research Networks]
o join in the User Services and associated WG sessions.
o co-chair a WG session on networked information retrieval tools.
o co-chair a BOF session on network training materials.
The following informal report is in note form and deals mainly with
the areas of User Support and Networked Information Retrieval,
although reports of the plenary sessions are also included. Whilst
it is as accurate as I can make it, it is naturally a personal account
and may be inaccurate due to lack of background information or
misinterpretation of what I heard. Corrections of fact are welcome,
but any discussion of items contained here would be best directed to
the appropriate mailing lists. Minutes of individual sessions are
also available via anonymous ftp from nri.reston.va.us
This report will be stored on the UK Mailbase Server. To retrieve a
copy, email to Mailbase@mailbase.ac.uk with the following command in
the body of the message:
send wg-isus ietf.11.92
Alternatively use anonymous ftp to: mailbase.ac.uk
file: pub/wg-isus/ietf.11.92
Note: in general I have not expanded acronyms as those readers
involved in a particular topic should know them whilst those who
aren't familiar with the acronyms should still be able to get a
reasonable overview of the topic.
All addresses quoted in the report are in internet (rather than UK)
order.
Each section has a double underlined heading - to enable you to skip
sections not of interest.
Plenary
=======
The plenary sessions and some of the working group (WG) sessions were
broadcast (audio and video) over the network. My impression was that
this did not go as well as it had done in Boston. There seemed to be
more problems - but this was almost certainly because they were being
more ambitious. There were many more sites involved and they were
attempting to broadcast two parallel sessions simultaneously.
Phil Gross welcomed attendees to the IETF and spoke about the problems
we're facing with regard to Internet Routing and Addressing. This was
the major technical issue at this IETF.
These problems are:
o scaling issues
o Class B exhaustion
o Routing table growth
o Overall IP address exhaustion.
The IESG (Internet Engineering Steering Group) had written an RFC
(Request For Comment) giving milestones for addressing the various
issues. The goal for the November IETF was to narrow the field of
IPv7 candidates if possible; to establish firm criteria; to give
specific feedback on each candidate based on the selection criteria.
Milestones:
By December 15:
Publish firm selection criteria as RFC and post constructive
feedback on candidates.
February 12:
At least two multiple interoperable implementations.
February 26:
Second draft of the protocol report (from each candidate)
indicating how each of these meets selection criteria.
There were technical presentations on the various contenders.
PIP (P Internet Protocol) using EIP (Extended Internet Protocol)
----------------------------------------------------------------
- by Paul Tsuchiya (Bellcore).
Paul felt that PIP was not mature enough to be chosen at this
time. It really needed a year of design/implementation and
testing. The question was "was it worth waiting for?". It is a
very general protocol that copes with all known
routing/addressing paradigms and is relatively efficient - and
more importantly - easy to evolve. Various sections of the
design had been been specified and there were already some simple
implementations of PIP hosts and routers.
PIP has been linked with EIPIP (see above). The PIP header would
be placed in what looks like an IP option field.
TUBA
----
Peter Ford (LANL) and Mark Knopper (MERIT).
TCP/UDP over CLNP addressed network.
The basic problem is that we're running out of IP address space.
We need to allow for >10**7 networks and >2**32 end systems. We
need more effective management of the IP address space to expand
the lifetime. (The CIDR addressing plans are already going into
effect.) TUBA is TCP and UDP with >32 bit addresses. This would
allow a hierarchical "sparsely populated" address space.
The TUBA transition will focus on the network layer. TCP and UDP
will be made to work over CLNP network layer. This will allow
the TUBA work to build on the existing CLNP experience.
SIP
---
Simple Internet Protocol - Steve Deering (Xerox Parc)
This is a new version of IP building on the lessons learned from
IPv4. SIP uses 64 bit hierarchical addresses. It is more
efficient than IPv4 in that it is only examined by identified
routers.
IP Address Encapsulation: Bob Hinden
------------------------------------
IPAE is intended as a transition mechanism for the new Internet
Protocol. They have been working with the SIP WG. Unlike with
TUBA - the host software does not need changing.
These issues are also discussed in Nov 1992 Vol 6 No. 11 of
"Connexions - The Interoperability Report" (ISSN 0894-5926)
(Email contact: connexions@interop.com)
NISI - Networked Information Services Infrastructure WG
=======================================================
Chair: Pat Smith (Merit)
This group is concerned with co-ordinating NICs (network information
centres) and improving the service they provide.
NIC Profile:
In order to aid NICs talking to each other, a NIC Profile
template had been produced. Chris Weider at Merit had set up a
server, which would accept completed NIC Profile Templates and
put them into X.500.
To find out how to submit a NIC Profile, NIC co-ordinators
should email to: X500test@merit.edu
Subject: help
They will receive the template and instructions on how to
complete it. The server auto-adds completed templates emailed to
it.
The Providers would be responsible for updating the information.
Several volunteers agreed to look at up-date mechanisms and to
decide on the intended audiences.
NETHELP:
The idea is that a user will type "nethelp" and be presented with
information on how to use the network. This would initially
simply be a pointer to local contacts and more information.
I again expressed my reservations. (Apart from the obvious
problem that the user has to know to type "nethelp" - not
terribly intuitive.) The group seemed to assume that the
technical people would be able to come up with a solution - so
long as NISI could specify what it wanted from nethelp. The
vague idea was that maybe this would be some software to be
installed on every PC or Mac etc (a daunting task in itself). I
felt we should be bringing the technical people into the
discussion and to try something. Ed Krol volunteered to knock up
a "straw person". This seemed a sensible course of action,
especially as NISI has been discussing nethelp for well over a
year.
Mailing list: nisi@merit.edu
To join, mail to nisi-request@merit.edu
Also:
Mailing list: nic-forum@merit.edu
To join, mail to nic-forum-request@merit.edu
The latter list is little used. Pat suggested it would be useful to
get all NIC personnel on to this list.
IAFA: Internet Anonymous File Archive WG
=========================================
Chairs: Peter Deutsch and Alan Emtage (BUNYIP)
This was probably the last meeting of this Working Group. The draft
version of the IAFA document(s):
Part I: A Guide to FTP site administration
Part II: Publishing Information on the Internet with anonymous ftp
had been discussed extensively on the IAFA mailing list. It was
agreed that the document should be put forward as a draft RFC. The
second of these documents suggests additional information that should
be made available on files (etc) in an archive (and about the archive
itself). The aim is to provide more information about the purpose and
content of files in an archive (other than simply the often
meaningless file name) for use by indexing tools such as archie.
It was agreed that in parallel several volunteers should look at
"iafa-ising" their file archives. It will only be when we try to put
these documents into practice that we will see where (if any) the
problems are. The next problem will then be convincing new archive
sites to provide the additional IAFA information about their archive
and individual files - and to add this information for the vast
existing base of file archives.
Simon Spero volunteered to add IAFA information to the SUN site file
archive (sunsite.unc.edu telnet to port 43).
Someone said they felt that the first IAFA document was "guidance
free" and suggested it needed to outline some "good things to do".
These would be recommendations for good practice rather than
requirements. Ellen Hoffman (Merit) said that when they put up their
anonymous ftp archive they looked around at several others and found
that some were definitely easier to navigate around than others. As a
result of this they have a check list of good points for archives
which they will make available.
April Marine (who was unable to attend due to lack of funding) had
written an introductory document to anonymous ftp. This had been
discussed on the mailing list and was well received. Other
suggestions for improvements were made at the meeting.
The IAFA document is available from:
archives.cc.mcgill.ca in the directory pub/iafa
Mailing list
iafa@cc.mcgill.ca
to join, mail to iafa-request@cc.mcgill.ca
POISED (Plenary)
================
POISED = Process Organisation for Internet Standards and Developments
Steve Crocker
This group was set up as a result of the Boston "unpleasantness" and
to answer the questions: who is in charge? who has the right to
determine policy and make decisions? etc. According to Steve, it had
resulted from the:
o IPv7 debate in particular
o Poised WG
o Real problems
o Scaling
o Delay in getting things done
o Communication difficulties
o Surprise
o Focus
o Burn out
Requests to join the mailing list for associated discussion to:
poised-request@cnri.reston.va.us
The WG will produce an internet draft.
There were to be three meetings - this plenary, a working group
session and the plenary on the last evening of the IETF. A
presentation to the Internet Society trustees was planned for December
10th.
Various proposals had been put forward. Maintaining the status quo
was an option. POISED Working Group was suggesting that a
redefinition and tightening up of the existing structure was now
needed. The IETF had grown to the point where a certain amount of
rethinking of the structure and procedures was required.
The draft proposal was for an Internet Technical Task Force (ITTF)
which integrated the IETF, IRTF and IESG back into one community.
The Working Groups would continue to remain the main focus and would
act as open forums. There would be different types of Working Group:
engineering, research and architectural. Working Groups would contain
self selecting design teams - which would go away and come up with
proposals and bring them back to the main group. This is recognising
what happens now in practice. The ITTF leadership would be provided
by a Technical Board (consisting of area chairs, ITTF Chair and
architects and performing the same role as the current IAB and IESG)
an Editor, Process Board and The Internet Society. Procedures for
selection of the officials and their acountability needed to be drawn
up and there needed to be mechanisms for removing members if
necessary.
The full details are in the POISED draft RFC. This needed further
discussion and if there was concensus in the IETF on it a transition
period would see the selection of the various officials.
User Services Working Group Chair: Joyce Reynolds
===========================
US-WG is the umbrella WG for the various user services area WGs. This
is the group which spawns new WGs and coordinates the work in this
area.
Mailing list for this group: us-wg@nnsc.nsf.net
To join, mail to: us-wg-request@nnsc.nsf.net
As an aside:
two new books written by IETF us-wg people had just hit the book
sellers and were very well received. Both are in a very readable
format.
Ed Krol:
The Whole Internet User Guide and Catalog
Tracey La Quey Parker:
The Internet Companion
the latter has a foreword by Al Gore.
Newcomers Guide
---------------
According to Steve Coya (IETF Secretariat), approximately 40% of
attendees at IETF meetings are first-timers. This means that for
a significant proportion, getting up to speed is a vital concern.
For the first time ever there was a Newcomer's orientation
session on the Sunday afternoon before the IETF. There had been
much discussion amongst "old" IETF-ers as to what should be
included and what needed documenting. Gary Malkin had written a
draft FYI RFC for newcomers. (Internet draft:
draft-malkin-newcomers-guide-00.txt) This document "Tao of the
IETF" was seen to be very useful - and was distributed at the
newcomers session. I attended the session out of interest (along
with about 100 others). The session was felt to be of use - but
was a little pedantic in parts. I felt there was too much
emphasis on how to register for an IETF (which most attendees had
managed to do!) and not enough on what the various Working Groups
were about.
This was borne out by the comments of the first-timers at the
US-WG meeting. It was suggested that the titles of Working
Groups were often not clear enough and that there should be short
descriptions of each group and what files were available and
where. This information would be best circulated prior to the
meeting to give people a chance to get up to speed. Perhaps a
pointer to "where to get information" could be given in the IETF
meeting announcement.
Steve Coya said they were considering doing a survey of the new
attendees and this would ask them what their expectations were,
etc. Comments on the newcomers session should be sent to:
ietf-info@cnri.reston.va.us
US-WG also suggested that a guide for new working group chairs
would be useful.
Internet "Quick and Dirty" Document: Peter Deutsch
--------------------------------------------------
Peter had posted a copy of his revised document to the us-wg on
the Saturday evening before the IETF. Most people had not had a
chance to read it. Suggested title: "A user's catalogue of
internet tools". It was currently 8 pages long. It should be
reduced to about 2 pages and should contain an
overview/introduction.
It would be split into:
Taxonomy:
- The basics (telnet, email, ftp)
- Mailing lists and Bulletin Boards (Listserv, Mailbase, usenet)
- Interactive Information Delivery (gopher and World Wide Web)
- Directory Services (WHOIS, X.500)
- Indexing Services (archie, WAIS, online-libraries)
Where to go next?
- References to other FYIs and books
The draft document may be retrieved via anonymous ftp from:
archives.cc.mcgill.ca
file: pub/uswg/quickanddirty.doc
Comments to:
Mailing list
us-wg@nnsc.nsf.net
and to join, mail to: us-wg-request@nnsc.nsf.net
RIPE Report: Joyce Reynolds
----------------------------
Joyce mentioned that she'd attended the last RIPE meeting
(Association of European IP Networks). She said that the Eastern
Europeans in particular are very keen to get all they can in the
way of RFCs and Internet Documentation. Some of the RFCs (in
particular the FYIs) had been translated into Czechoslovakian.
Eastern Europe
--------------
Alan Emtage (bajan@bunyip.com) talked about the initiative to
send unwanted old editions of computing and networking text books
to eastern Europe. The Internet Society is acting as a
repository for donations. Alan suggested that individual sites
might like to act as network "buddies" to eastern European or 3rd
World sites which are just embarking on networking.
RARE ISUS WG Report: Jill Foster
--------------------------------
The RARE Information Services and User Support WG met in Pisa at
the beginning of November, immediately preceding the European
Network Services Conference. An "electronic meeting" had also
been held in September. I had previously circulated a report to
the us-wg mailing list prior to the IETF. The main points
covered were:
o formation of joint IETF/RARE/CNI WG on networked information
retrieval (see later in this report).
[CNI - Coalition for Networked Information]
o co-ordination and registration of European gophers.
o formation of joint IETF/RARE network training materials
group.
o starting to get librarians involved (very important -
particularly in NIR work).
o need for a 'Total solution user information' - access to all
network services via one GUI as the ultimate goal.
o update (Nov) of RARE Technical Report 1 on information
services and user support in RARE community
[available via anonymous ftp from: mailbase.ac.uk
file: pub/rare-wg3-usis/rtr-usis-92 ]
There is a need for collaboration and co-operation on user
support in Europe and the world. Users can roam the global
network so we have
"Le 'user' sans frontieres"
What we need now is
"Le 'user support' sans frontieres"
i.e. co-operation on a global scale without worrying about
whether we're members of RIPE, EARN, RARE, IETF, etc.
Mailing list of ISUS: wg-isus@rare.nl
To join, mail to: mailserver@rare.nl
the command (in the text of the message):
subscribe wg-isus firstname lastname
(substituting your own first and last name)
FYI 4 and 7 (FAQ): Gary Malkin
------------------------------
FYI 7 "FYI on Questions and Answers: Answers to Commonly Asked
'Experienced Internet User' Questions", (Also RFC 1207), February
1991.
FYI 4 "FYI on Questions and Answers: Answers to Commonly Asked
'New Internet User' Questions", (Also RFC 1325), May 1992.
Gary wanted to update the new user and experienced user
Frequently Asked Questions (FAQ) RFCs. He suggested "creating"
questions for FYI 4.
FIY 7 contains actual questions.
FYI 4 is RFC 1177, 1206, 1325: The latest version being RFC 1325
- May 1992.
Now need to look at experienced user FAQ (FYI 7). He wanted
volunteers to monitor lists and mail in FAQs to quail discussion
list.
To join the list, mail to: QUAIL-request@XYLOGICS.COM
It was suggested that people who had to answer user questions
should send in their own "top 5" frequently asked questions.
Document: What is the Internet? Ellen Hoffman
----------------------------------------------
Ed Krol had had permission from his publisher to put this
together from his book. The draft document was circulated for
comment.
RFC 1359/FYI16: Connecting to the Internet
------------------------------------------
FYI 16 "Connecting to the Internet: What Connecting Institutions
Should Anticipate", (Also RFC 1359), August 1992.
This RFC explains what is required to get connected to the
Internet and whom to contact. Joyce felt it needed to be
expanded and should move away from being US centric. Pat Smith
said that the RFC is a good start but that it currently aims at
the university community and the Internet is now broader than
this. The RFC needs separate sections for universities, schools,
individuals at home etc. European input from RIPE and RARE was
requested.
User Document 2: Ellen Hoffman and Lenore Jackson
=================================================
FYI on Introducing the Internet
-------------------------------
This was a short bibliography for a complete newcomer on where to
get more information on getting connected, etc. The draft will
be revised following comments and then recirculated on the user
doc mailing list.
Comments on Documentation
-------------------------
In the past US-WG has kept away from providing end user
documentation. Ed Krol said we should stop writing technical
documents supposedly for end users. It was suggested (by someone
else) that US-WG should not in fact be writing documents for end
users - we should leave that to those in the community who write
professionally.
The fact is that over the last year we have seen documents being
produced that are being used by real users. They were all they
had - and so better than nothing. There are some good books now
written - but it still remains a worthwhile effort to write short
end-user oriented jargon-free documents that are freely available
- and which could be taken and incorporated into a site's set of
end user documents. I personally feel that us-wg should not
spend effort in typesetting documentation (as discussed) - but
should spend time collecting good short descriptions in ascii
text - that a site can take and format in its own house style.
We've only made a start in this area and have a long way to go
before the documents produced are suitable for end users. Some
of the documents overlap and the content and audience are not
always obvious from the title. Some rationalisation is required
here. Meanwhile sites such as Merit (on nis.nsf.net) are setting
up directories called
introducing.the.internet
with some of these articles which end users are consuming
hungrily.
Mailing list: user-doc@merit.edu
To join the list, mail to: user-doc-request@merit.edu
Information Infrastructure: lunch meeting
=========================================
The chairs of the Working Groups concerned with various aspects of
information retrieval and tools (IIIR, NIR, WNILS, URI, URL) met over
lunch to discuss co-ordination of the work of these groups. A
statement is to be drawn up to show how these groups inter-relate and
what parts of the puzzle they are intended to cover. Note that the
NIR Group was set up explicitly to help track the various different
groups and projects in this area. It was decided to draw up a list of
areas, so that if for example a new tool was developed which
effectively indexed information resources - then it would be dealt
with by a generic "indexing tool working group" rather than set up a
new working group for that application.
I personally feel that there needs to be some tightening up of the
descriptions and scope of the various working groups. It is not at
all obvious to new members (or old!) as to which group deals with
what. Some of the descriptions are very similar and some quite
misleading. For example: NISI - Network Information Services
Infrastructure sounds like it should be concerned with information
architecture - whereas it was set up primarily to improve the dialogue
between NIC personnel.
Training Materials BOF: Ellen Hoffman and Jill Foster
=====================================================
The proposal was to form a joint IETF/RARE Working Group on network
training materials and the problems associated with network training.
The main objectives are:
o to provide a comprehensive package of "mix and match" training
materials for the broad academic community.
o to provide a catalogue of existing training materials.
I briefly described the UK NISP/ITTI Training Materials Project (which
has very similar aims) and circulated a print out of the catalogue of
training materials collected so far. The fields used were based on
the TopNode Data Elements - but some more discussion is needed. (Pete
Percival and Craig Summerhill of TopNode were at the IETF.)
The members of the BOF (Birds Of a Feather) introduced themselves and
described their training activities. Many gave presentations to their
users on using the network and the networked services, but only a few
provided hands on training. Those that were noted that this was very
popular and successful. The group agreed to help to produce a
catalogue of training materials on using the network (using the
TopNode Data elements). CNIDR volunteered to host that database. The
group plan to evaluate existing training materials and to identify
materials to be produced. Problems related to network training were
also discussed. The charter was agreed (with some changes for
clarification) and it was proposed that a Training Materials WG be
formed.
Discussion is to take place initially on the us-wg list. Other
"network training" lists are:
o the "Internet/BITNET Network Trainers" list
NETTRAIN@ubvm.cc.buffalo.edu
(which is quite "noisy" but nontheless useful)
To join, mail to: listserv@ubvm.cc.buffalo.edu
the command (in the text of the message):
subscribe nettrain firstname lastname
(substituting your own first and last name)
o the UK ITTI Network Training Materials Project email list
itti-networks@mailbase.ac.uk
To join, mail to: mailbase@mailbase.ac.uk
the command (in the text of the message):
subscribe itti-networks firstname lastname
(substituting your own first and last name)
The Merit/NSFnet Training Sessions included presentations followed by
informal hands on workshops. Their Internet Cruise (disk based
presentation) was also well known. It had been translated into
Spanish:
"El Cruso!"
Also mentioned were:
o The Internet Hunt (a monthly set of questions, the answers to
which can be found scattered across the Internet)
o Mining the Internet: UC Davis - a hands on workshop
o Navigating the Internet. An Interactive Workshop. (Made
available over the network and which closed subscriptions at
15,000!).
Some sites put information on disk for training session attendees to
take away. This information ranged from: Zen and the Art of the
Internet to the various directories of information and mailing lists.
Networked Information Retrieval WG
==================================
Co-chairs: Jill Foster, George Brett
This is a joint working group between the IETF, RARE ISUS WG and CNI.
George Brett reported on the $1.2M funding from NSF for a new Clearing
House for Networked Information Discovery and Retrieval (CNIDR).
George Brett is Director, Jane Smith is Assistant Director. Jim
Fullton will be working as their programmer. Given the current
concerns about the reliance of the community on software tools
developed by volunteer effort and the pressure being brought to bear
on some of these volunteers by their employers - the setting up of a
centre for support was welcomed.
I reported on the RARE ISUS WG meeting and the sessions concerned with
NIR. The Pisa conference had sessions on WAIS, WWW, gopher, archie,
Soft Pages and Hyper-G. ISUS WG members have agreed to co-ordinate
informally the European gophers and to work on subject specific
gophers. This work and discussion is progressing in co-operation with
the IETF/CNI groups and TopNode Project, Library of Congress and some
Australian gopher people.
Since the Boston NIR BOF, the templates, for collecting information
about groups involved in NIR work and about the various NIR tools and
applications, had been agreed on the nir mailing list. The
information on most of the groups and tools had been collected from
the appropriate contacts and had been edited together into a draft
document which had been made available prior to the IETF.
Access details: anonymous ftp from mailbase.ac.uk
file name: pub/nir/consumer.report
At the meeting the draft of the "Consumer Report" on NIR Tools was
reviewed and discussed at length. There was discussion about the
tools and the groups that have been included in the report. It was
suggested that Hyper-G and Soft Pages should not be in the main part
of the report. An appendix of "Forthcoming Attractions" was suggested
and it was felt that a short description of Hyper-G and Soft Pages and
WHOIS++ would be appropriate there. A number of other appendices were
recommended to provide a broader base of information for the users and
providers of the information. The templates and the report were well
received. It is planned to update the templates on a quarterly basis.
The working group also discussed future plans to evaluate the tools,
and the general usefulness of this report. This "evaluation" was seen
more as completing a matrix of characteristics on each tool - so that
the "consumer" could check for the features required and see which
tools provided them. The evaluation would therefore be an objective
checklist rather than of the "best buy" variety. The working group
concluded its work with a discussion about future dissemination of the
report in varied media. There was consensus that the report will have
to be made available across the many networked information retrieval
tools and that we will have to be prepared to ensure that the form of
the information accommodates these tools. One of the first online
archives for the "Consumer Report" on NIR Tools will be hosted at
CNIDR. The working group will continue to collect and verify
information for the templates.
The nir list is currently hosted at McGill and needs to be moved. All
subscribers should now subscribe to the list on the UK Mailbase server.
list: nir@mailbase.ac.uk
to join: mail to: mailbase@mailbase.ac.uk
the command (in the text of the message):
subscribe nir firstname lastname
(substituting your own first and last name)
For example: subscribe nir Jill Foster
archives in directory: pub/nir on mailbase.ac.uk
An announcement will also be made on the existing nir list.
Uniform Resource Indicator WG: Peter Deutsch, Alan Emtage
=========================================================
URI is now the union of Uniform Resource Location and Uniform Resource
Indicator. The idea is to identify information resources uniquely and
to allow the location of these by navigational tools.
Tim Berners Lee (who was unable to attend due to lack of funding etc)
had adapted one of his WWW documents to discuss URIs. There was some
discussion of this (which was a little one sided as Tim wasn't there
to defend it). There was some objection to the use of "name" in the
document and some discussion on "fragmentation" and whether to allow
partial form URIs.
The group felt the document should have an overview section that could
be read on its own. This should define the various acronyms (URI,
URL, URN, URSN, etc). It was felt that there should be the
Berners-Lee document and then a document putting the counter
arguments. Input from the library community is needed and the
overview document will be circulated widely - including to the CNI.
mailing list: ietf-url@merit.edu
to join, mail to: ietf-url-request@merit.edu
[I'm not sure which list is currently being used - but mail to
ietf-url-request and ask to be put on the list for URI discussion!]
WHOIS and Network Information Look Up Service Working Group: WNILS:
======================================================================
Joan Gargano
This group had met as a BOF at the last IETF. It had been a stormy
meeting with the X.500 set saying why re-implement X.500 features in
WHOIS when X.500 does it all and the WHOIS set saying - "why not?
Maybe we can do it quicker and with less overheads."
Since the last IETF Jim Fullton, Peter Deutsch, Chris Weider and
others had worked on implementing a WHOIS++ Server and had some
prototype software working.
The WHOIS++ project aims to develop a lightweight useful Internet
Directory Service using simple technology.
Features
o template oriented data model
o simple command syntax
o database technology selected by operator
o "centroids" system allows summary information to be easily
propagated
Advantages
o Easy to use
o Servers easy to create, install and maintain
o database maintenance separate from server (allows using existing
database)
o clients easy to write
o centroids simplify the task of finding suitable servers
Structure of WHOIS++ database (logical)
Template type 1 type 2 type 3 etc
--------- ---------- ----------
| 1 | | | | |
--------- | ---------- | ---------- |
| 2 | | | | | | | |
--------- |-- ----------- |-- ----------- |--
| 3 | | | | | | | |
| |-- | |- | |-
| | | | | |
--------- ----------- -----------
Every item and each database has a unique WHOIS database handle
(issued by IANA).
Search terms: "handle", template type, attribute.
The work the sub-group was doing was a proof of concept piece of
work.
Architecture:
Client ---> Front End
|
V
Search Engine ----> optional gateway
\
\
--> database
Centroid:
Centroid information propagates up the tree. Clients can query parent
servers to find servers with given keywords.
Keywords stored only once and just hold pointers. Can search at any
level. Bottom up driven approach to finding information.
Command Set:
o Template oriented
o backwards compatible with WHOIS
o system commands and search commands
o search command: Attribute, Value, Template or Handle, Search All
o Constraints on searches
Status since Boston:
o Bar BOF in Boston
o Joan Gargano's overview document
o Peter Deutsch's basic architecture
o Chris Weider etc centroids architecture
(sent to list - are in archive)
o WNILS working group now formed
o Centroids propagating: 2 running now, more to come.
o Public domain version. Then supported by BUNYIP. Clients -
X-Windows, etc.
o Centroid code written
Chris Weider:
Centroids: "forward" information propagated around.
A poll mechanism allows the server to get forward information
from below. Servers know what it is being polled by and can let
server "above" know when information is changed. So a user can
query any level in the mesh and the query is passed to the point
in the tree where it is most likely to be answered. Don't need
to know where or how servers/data distributed in order to be able
to prune search tree and find information want/looking for. With
a distributed mesh it is easy to build Yellow Pages services. A
server can specialise on a particular attribute.
Having got a rough prototype - what's next?
They talked about privacy and security issues, replication of servers
etc. There are currently 100 WHOIS servers which don't talk to each
other. WHOIS++ would help to provide some glue between their servers
and encourage new servers.
It started to sound all too familiar: Problems of replication and
security; The massive problem of getting the data in the database -
All problems X.500 has been faced with. We'll have to see whether
WHOIS++ fares any better. Peter Deutsch likes to talk about Darwinian
selection!
Integration of Internet Information Resources Working Group (IIIR):
======================================================================
Chris Weider
The purpose of this working group was to start to pull together some
of the applications (WAIS, gopher, archie) and to work on
interoperability issues, what new tools should do and to discuss
gateway protocol design.
Several people were asked to report on the work they were doing in
this area.
Tim Howes and gopher/X.500 gateway:
This was proving popular with some 2000 queries/day to the
UMichigan directory via the gopher/X.500 gateway. (This is
2000/day out of 15000/day total queries at U Michigan). Various
other sites are running such gateways too.
Tim said people don't want WAIS, archie, gopher or X.500 - they
want information and they don't want to flip clients to get
different information. [Agreed! Neither do they want to swop
tools to change from searching to browsing.]
Peter Deutsch: WHOIS++
Peter talked about WHOIS++ (see above) and the possible use of
Prospero to provide user centred views of the information (as
Prospero currently does for archie).
Jim Fullton: WAIS
Jim had been working with NASA. They have satellite data
accessible via WAIS. A user can choose the format for display.
Gopher has no concept of different formats. The main problem was
gatewaying from the "maximal" WAIS server to the "minimal" gopher
world.
The group decided to write an Internet draft on: Gateway Protocol,
common exchange format and query routing protocol. A registry of
gateways would be useful too.
We need to avoid an explosion of protocols that are basically the
same. (One of the aims of the NIR report is to help contain this
explosion by disseminating information about what is already
available.) We need "classes" of protocols - with one working group
per class. It was stated that we're not a closed community and that
we should bring in the information retrieval people, the IRTF group on
information retrieval (Mike Schwartz etc) and the librarians.
A mailing list was not mentioned for this group. However enquiries
could be sent to ietf-url-request@merit.edu
Concluding Remarks
==================
This IETF seemed more hectic than normal with more Working Groups and
meetings over dinner and lunch. There was unfortunately little time
for general conversations or (for me) to attend the Mail and
Directories Working Group meetings. Shortly before the IETF - the
closure of the Automatic Mailing List Server WG was announced. I had
understood it had been put on hold, so was disappointed to hear this.
On a positive note: the CNI held their meetings near Washington at the
end of the same week as the IETF. Several key CNI members attended
the IETF and there was to be cross representation at the CNI meetings
too. George Brett and Alan Emtage had set up meetings at the Library
of Congress. OCLC and the Library of Congress are working on
classifying or cataloguing networked information resources. The
TopNode Project is looking at doing this too - but are, I believe,
taking a different approach. This was an opportunity for these people
to get together and to start mutual discussions.
So - I feel that this IETF and the events surrounding it saw another
significant milestone on the road to bringing order to the chaos as
far as networked information is concerned. We still have a long way
to go in satisfying end users of course. We still have the plethora
of different tools - but at least we're all starting to pull in
approximately the same direction.
Finally, a reminder that these notes are my view of the IETF. They
may not be an accurate view, and certainly do not cover the wide range
of topics discussed at the workshop. Apologies for the delay in
getting this report out. I went straight from the IETF to Australia
to take part in their Networkshop and to look at some of their network
training, returning shortly before Christmas; but that's another
story....
Jill Foster (Jill.Foster@newcastle.ac.uk)
28.12.92
>From nir-request@kona.cc.mcgill.ca Tue Jan 5 11:51:34 1993
Received: from kona.CC.McGill.CA by othello.admin.kth.se (5.65+bind 1.8+ida 1.4.2/4.0b)
id AA28539; Tue, 5 Jan 93 11:51:32 +0100
Received: by kona.cc.mcgill.ca (5.65a/IDA-1.4.2b/CC-Guru-2b)
id AA24001 on Tue, 5 Jan 93 04:41:40 -0500
Received: from cheviot.ncl.ac.uk by kona.cc.mcgill.ca with SMTP (5.65a/IDA-1.4.2b/CC-Guru-2b)
id AA23997 (mail destined for /usr/lib/sendmail -odq -oi -fnir-request nir-out) on Tue, 5 Jan 93 04:41:36 -0500
Received: from eata.ncl.ac.uk by cheviot.ncl.ac.uk id <AA29650@cheviot.ncl.ac.uk>
(5.65cVUW/NCL-CMA.1.35 for <nir%cc.mcgill.ca@newcastle.ac.uk>) with SMTP; Tue, 5 Jan 1993 09:41:33 GMT
Date: Tue, 05 Jan 93 09:41:30 GMT
From: Jill.Foster@newcastle.ac.uk
Subject: NIR list moving
To: nir@cc.mcgill.ca
Cc: nir@mailbase.ac.uk
Message-Id: <emu-ct08.1993.0105.094130.njf@eata.ncl.ac.uk>
Status: RO
Attention all list members!!
Way back in November - Alan Emtage, Peter Deutsch and I discussed
moving the nir list from McGill to Mailbase at Newcastle. This was
agreed at the IETF meeting in Washington. (I've been on the road or
on holiday since - so have only just got back to it.)
The McGill list should be used until an announcement that the new list
is operational.
The reasons for moving;
o McGill may not be willing to host the list now that Alan and
Peter have "moved" to BUNYIP.
o Members had to be added to the list by hand and therefore when
Alan or Peter were busy/travelling this could cause delays.
o There was no way of seeing who else was on the list.
The list is now "global" - so hosting it in the UK rather than Canada
isn't a problem.
Mailbase is an automatic mailing list server. It requires your name
as well as your email address (taken from your message header) in
order to produce a meaningful list of members. As the McGill list
does not have your names (only email addresses) - you will need to
subscribe to the new list. Apologies for the hassle!
to join: mail to: mailbase@mailbase.ac.uk
the command (as the only text of the message):
subscribe nir firstname lastname
(substituting your own first and last name)
For example: subscribe nir Jill Foster
Message archives and files associated with nir will be in directory:
pub/nir on mailbase.ac.uk
They may be retrieved via email or anonymous ftp.
Please subscribe now while you remember.
Mail for NIR should be sent to the McGill list until you see an
announcement on the two lists saying that the move is complete.
Mail to the new list should then be sent to: nir@mailbase.ac.uk
Thanks for your patience and cooperation.
-- Jill Foster (NIR WG chair)
>From nir-request@kona.cc.mcgill.ca Tue Jan 5 21:00:54 1993
Received: from kona.CC.McGill.CA by othello.admin.kth.se (5.65+bind 1.8+ida 1.4.2/4.0b)
id AA10382; Tue, 5 Jan 93 21:00:47 +0100
Received: by kona.cc.mcgill.ca (5.65a/IDA-1.4.2b/CC-Guru-2b)
id AA27629 on Tue, 5 Jan 93 12:24:13 -0500
Received: from att-out.att.com by kona.cc.mcgill.ca with SMTP (5.65a/IDA-1.4.2b/CC-Guru-2b)
id AA27623 (mail destined for /usr/lib/sendmail -odq -oi -fnir-request nir-out) on Tue, 5 Jan 93 12:23:57 -0500
Message-Id: <9301051723.AA27623@kona.cc.mcgill.ca>
From: egrimmelmann@attmail.com
Date: 5 Jan 93 16:51:16 GMT
To: com-priv@uunet.uu.net, disi@merit.edu, members@farnet.org,
ietf@cnri.reston.va.us, nir@cc.mcgill.ca, nisi@merit.edu,
osi-ds@cs.ucl.as.uk, wais-talk@think.com
Cc: london@attmail.com (Shelly London ),
estradas@nic.cerf.net (Susan Estrada ),
scottw@nic.ddn.mil (Scott Williamson ),
dmitchel@nsf.gov (Don Mitchel ), steve@nsf.gov (Steve Wolff )
Received: from egrimmelmann by attmail; Tue Jan 5 17:16:05 GMT 1993
Phone: 908-234-6798
Fax-Phone: +1 908 234 7655
Subject: NSF Directory/Database Award
Content-Type: Text
Status: RO
We apologize if you receive multiple copies of this announcement; it is posted
to several news groups
======================================================
AT&T PRESS RELEASE
For further information:
Shelly London 908 221-4355
FOR RELEASE JANUARY 5, 1993
AT&T Will Provide Directory & Database Services to the National Science
Foundation NSFNET
Basking Ridge, NJ--AT&T announced today that it has signed a cooperative
agreement with the National Science Foundation to provide directory and
database services for NSFNET, the National Science Foundation national data
network that is part of the Internet. The Internet is comprised of more
than 5,000 computer networks that facilitate collaboration among members
of the research and education community. The Internet, and in particular
NSFNET, is projected to serve as a basis for evolution to the National
Research and Education Network (NREN).
Under the terms of the agreement, AT&T will develop and maintain a
Directory of Directories which will serve as a pointer to numerous
resources on the Internet. It will include lists of FTP (File Transfer
Protocol) sites, lists of various types of servers available on the
Internet, lists of white and yellow pages directories, library catalogs
and data archives. The Directory of Directories will enable even novice
users to obtain references to information they need through simple,
easy to use interfaces. AT&T also will provide white and yellow pages
type directory services, such as names of users, organizations and
resources on the Internet, using X.500 technology, the current standard
specification for distributed information storage and retrieval.
As part of its database services, AT&T will establish database servers to
extend and supplement the resources of the NSFNET, including databases of
contributed materials of common interest to the user community and
communications documents. AT&T also will offer database design,
management and maintenance services to organizations and groups for
inclusion in the Internet.
Initially, access to all services will be provided through several
currently popular in-use interface methods; with time, it is anticipated
that X.500 will become the primary method of access.
In providing these services, AT&T will work cooperatively with two other
organizations: CERFNet, a General Atomics project, which was awarded a
similar agreement for information services, and with Network Solutions,
Inc.(NSI), which was awarded a similar agreement for registration
services. The three corporations will collaborate under a common
concept called INTERNIC.
"We all feel intuitively that the domestic Internet and the distributed
collaboration that it facilitates are rapidly creating a national
'workplace without walls'", said Steve Wolff, Director, Division of
Networking and Communications Research and Infrastructure, NSF. "These
three awards to geographically dispersed organizations for Network
Information Services will both exploit and demonstrate the success of the
network in enabling distributed collaboration."
"These directory and database services are essential components of the
emerging national information infrastructure," said Erik Grimmelmann,
Marketing Director, Internet/NREN, AT&T Data Communications Services.
"This agreement marks an important step for the Internet as well as for
AT&T because services such as these and the related ones to be provided
by our INTERNIC collaborators will make the Internet even more useful
than it is today."
The cooperative agreement is for a five-year period, with annual reviews.
It is expected that the NSF will contribute approximately one third of the
costs, with another third provided by AT&T and the remainder recovered in
user fees. The user fees, which have been proposed for maintenance of
special databases and extensive directory listings, are consistent with
Federal Networking Council (FNC) cost recovery guidelines. The user fees
were part of AT&T's proposal, which was evaluated by an NSF review panel
and approved by the NSF. The full text of the NSF statement on INTERNIC
user fees is included at the end of this release.
The agreement is a natural extension of AT&T's strong commitment to
education, research and the advancement of high-speed data networking.
For example, AT&T operates XUNET (Experimental University Network), a
high speed experimental research network for the academic community, and
is a key participant in the CNRI (Corporation for National Research
Initiative) sponsored BLANCA gigabit testbed. AT&T also supports
collaborative applications research projects of direct relevance to the
Internet, including an information retrieval service, an image retrieval
service and a newly developed directory concept called "nomenclator" that
has been shown to improve response time tremendously when searching large
directories.
##
Text of NSF statement on user fees:
Consistent with FNC guidelines on obtaining reasonable cost recovery from
users of NREN networks, the NSF has determined that the INTERNIC
Information Services provider may charge users beyond the U.S. research
and education community for any services provided. Also, the INTERNIC
Directory and Database Services provider may charge a fee for maintenance
of special databases, for extensive directory listings and may charge
users beyond the U.S. research and education community. Finally, because
the registration function provided by the INTERNIC Registration Services
applies to domestic and international, commercial and individual users in
addition to research and educational users, it is expected that an
appropriate registration fee structure will take time to develop.
NSF expects to engage in an extensive discussion with the domestic and
international Internet community on the motivation, strategy and tactics
of imposing fees for these services during the next fifteen months.
Decisions will be implemented only after they have been announced in
advance and an opportunity given for additional public comment.
5