Path: utzoo!utgpu!watmath!uunet!bionet!irlearn.bitnet!NOREILLY From: NOREILLY@irlearn.bitnet ("Niall O'Reilly ", NOREILLY@IRLEARN.UCD.IE) Newsgroups: bionet.general Subject: EARN 89 Message-ID: <8906191037.AA28176@net.bio.net> Date: 19 Jun 89 10:36:52 GMT Sender: daemon@NET.BIO.NET Reply-To: Jill.Foster@newcastle.ac.uk Lines: 374 Jill Foster's account of EARN 89 ... Niall ----------------------------Original message---------------------------- EARN 89 - a Trip Report ======================= The following is a personal view of EARN 89. These notes are correct to the best of my knowledge, but there may well be points I have misunderstood. EARN 89 was held in May in Aghia Pelaghia in Crete, Greece; both the programme and location promised once again to be extremely interesting. One of the highlights of the conference was a trip to Knossos, followed by a Greek barbeque. EARN 88 in Turkey had a bias towards towards User Support and Applications, and EARN 89 continued this theme with special emphasis on applications in the Medical field. The programme consisted of parallel sessions on Medical Applications and on Networking (including network servers, databases etc.). Some participants I talked to were disappointed in the programme and felt that there was not enough technical content. My interests lie in networking (wide area and local area), Information Services and use of networks by Medics. I therefore found it extremely difficult to choose which parallel session to go to as they all promised to be interesting. The proceedings of the whole conference were made available beforehand. This was much appreciated, and helped in deciding between parallel sessions. The programme stretched from 8:30am to 7pm each day with a respectably long gap in the middle to allow mad dogs and englishmen to enjoy the mid-day sun and the beach. Internetworking --------------- The first day consisted of tutorials. Larry Landweber of Wisconsin gave a detailed session (four to five hours) on Internetworking. He concentrated on routers and bridges. One of the things that concerns me and many others is the different migration paths to OSI that are being taken by North America and Europe (US-> TP4 (Transport Service Class 4) and Europe -> TP0 (Transport Service Class 0)). I waited eagerly for the bit on bridging at the Transport Level. Unfortunately by the time Larry reached the Transport Layer, the session time had been overrun. He said that no-one in their right minds would consider bridging at the Transport layer. He didn't really expand upon this statement, but did say that a student of his had written a Transport Level gateway and had found various problems in trying to relay at this level. I wasn't alone in finding this statement a little radical, but after 5 hours and a coffeeless coffee break I hadn't the heart to prolong the session. (I did leave my email address for the report on the student project.) Larry made available a complete set of his tutorial slides, which was very useful. Strictly speaking relays at the Transport Level are against the ISO Model, but there is currently some input from working parties on the need to change the model in this respect. At the RARE (Association for Academic and Research Networking in Europe) European Networkshop in Trieste at the beginning of May, Transport Level bridging was being put forward as the solution to the US v Europe Migration problems. In fact Christian Huitema of Inria (France) has been instrumental in producing a TP0 to TP4 gateway. EARN President's Report ----------------------- Frode Greisen (the current President of EARN) gave a status report. EARN now serves 22 countries and has 700 nodes and 70,000 users. It encompasses machines from 30 different manufacturers, and 12 different operating systems. New connections during 1989 will include Algeria, Cyprus, Morocco and Tunisia. OSI Migration ------------- Frode Greisen gave an overview and Dennis Jennings gave a more detailed account of the plans for EARN to migrate to using OSI. The Migration is proceeding both top down and bottom up. Top down: Applications will be introduced as they become available: X.400, X.500, FTAM. Bottom up: X.25 Infrastructure is being put in place. Digital, Nothern Telecom and IBM have each donated equipment for this international X.25 infrastructure. The idea is to to run EARN protocols over ISO Session Layer over X.25. There is currently an experimental backbone network between Geneva, London, Amsterdam and Montpellier. The interworking between the various so called G-boxes (DEC) and E-boxes (IBM) is being tested. There should be an operational backbone by early July '89. Mention was made of IXI, the International X.25 Infrastructure proposed by RARE and COSINE (Cooperation for OSI Networking in Europe). The EARN BOD are curently investigating whether EARN could use this backbone. Aspects of Scientific Networking in Europe ------------------------------------------ Klaus Ullmann of DFN (German Research Network) and the President of RARE stressed the need for our currently fragmented European research community to present a united front to external bodies such as the PTTs and the funding bodies. He maintained that the various networks were all striving towards similar goals, and urged that the larger network organisations ask their technical people to cooperate. Having just attended the RARE conference, I was very aware that the same themes and questions were appearing at EARN 89 as had appeared at the RARE networkshop: OSI Migration and the problems that might be caused by the different migration stategies of Europe and North America. The need for enhanced user support and support for special interest groups. The need for publicity to make more researchers aware of the possible uses of the network to enhance their research. The need for cooperation between Europe and North American networking people, both at the technical level and at the user support level. The problems, particularly in Europe, of dealing with the various PTTs. From X.25 to ISDN ----------------- Louis Pouzin gave an entertaining talk on the "Changing Data Communications Scene". He described the highly successful Teletel service in France. This has proved to be extremely popular in the private sector, but professional use has been slow to get started. The French PTT gave each phone subscriber a free Minitel (videotex) terminal with which to access the service (in conjunction with their phone). He also described MDNS (Managed data Network Service). This was the original name proposed for IXI (the European X.25 Infrastrucure) but has been adopted by the European Company formed by the various PTTs. One of the problems in the past with European communication has been that there was no overall management. If there was a fault with an international connection, it was difficult to determine which PTT was responsible. MDNS will provide an X.25 service from the 21 European carriers. The idea is to provide 'one-stop shopping' and centralised network supervision. Other Networks -------------- There were various talks on other networks: EUNET: Daniel Karrenberg North American Networks: Larry Landweber HEPNET: R.Blokzijl Medical Doctors' and Researchers' Attitudes as Users of EARN ------------------------------------------------------------ Jean-Claude Salomon discussed some of the problems he had encountered in trying to set up a mailing list for those involved in cancer research. The list (CLAN - Cancer Liaison and Action Network) had seen little activity: many people signing on and subsequently signing off without participating. He felt that one of the main factors which inhibited the sharing of information and results was the importance researchers attached to the number of papers they published. Parallel Sessions ----------------- The rest of the conference consisted of parallel sessions on: A1: Existing Networks B1: Network Servers/User Interfaces A2&3: Medical Applications (I and II) B2: Network Monitoring, Management and Security B3: Access to Databases/Libraries. Below are brief reports on some of the sessions I attended (swapping between parallel sessions). A1: GROUPTALK - Jack Williford of IBM talked about the IBM product GroupTalk, which provides the user with a local interface to Conferencing Systems, Bulletin Boards and other related applications. B1: LIFESCI - A tool for Computer Supported Cooperative Work. Ami Zakai of Technion described LIFESCI which supports real-time conferencing, bulletin boards, electronic journals, application servers etc for those in the Life Sciences. It is based on the LISTSERV and Relay functions of EARN but has an enhanced user interface to allow users to enter 'english-like' commands. This is the sort of initiative that should be encouraged on our networks, as they become true academic and research networks, used by researchers from all disciplines. Our new users will not put up with the type of interface to servers that networkers and computer scientists tolerate. They will just not bother to use them. TRICKLE SERVER - Turgut Kalfaoglu described this server which was developed as a result of the need (identified at EARN 88) for a public domain software server on EARN. There are now 7 Trickle Servers in Europe which provide a caching service for files retrieved from the public domain software servers in the States. A file is only retrieved from the servers in the US if it is not cached on one of the Trickle servers in Europe. MAIL FILTER - Mitch Wyle (ETH Zurich) described a rule based Mail Filter that he had developed. This Mail Filter performed user-defined actions on electronic mail messages according to user-defined rules. The Filter sat between the Message Transfer Agent and the User Agent and preprocessed the mail before it was placed into the user's mailbox. This preprocessing happened automatically without requiring action from the user. GRASP INFORMATION SERVER - Jim Sweeton (NSFNET/Merit). Jim described GRASP (GRASP = GRAND + SPIRES) which provides computer facilitated discussion, file server functions etc. GRASP combines a front-end query resolver with a database management system as the back-end search and report facility. The services are available interactively, by FTP and via mail based queries. The Information Services are primarily to support the NSF associated network operations and information centres, but also support the end users on the Mid-level networks attached to the NSFNET backbone. The file server and distribution-list support is very like that of LISTSERV. In the medium term they are working on a local user interface program which will run under MS/DOS or on the Mac and which will communicate directly with the server. Long term goals were to establish a distributed information system in cooperation with the other network information centres. There should be standards for information storage and retrieval across networks. Apparently there is an RFC on Information Services which is due out in the near future. There is a Network Information Service Infrastructure Working Group in the US. This is a subgroup of IETF User Services Working Group . They will be producing a white paper on recommendations for NIC interoperability. The US NICs (Network Information Centres) involved so far are: BITNIC, SRI-NIC, CSNET, Merit/NSFnet, NNSC. I'm a member of a subgroup of RARE Working Group 3 (User Support) which is concerned with Information Services. I've asked Jim to keep me in touch with the activities of the IETF subgroup. A2: TRAINS - Eric Keppel described the project which uses EARN to transfer information associated with kidney and heart transplants. A database is kept of patients requiring transplants. Detailed information on factors affecting rejection of transplanted organs are kept. There are very few donor organs and a very long waiting list for them. It is essential to be able to optimise the matching of a patient to an organ so as to mimimise the risk of rejection. This needs to be performed very quickly of course, and EARN has been used very successfully in this field. The main problem encountered (which was echoed by others talking on medical applications) was the fact that EARN access is often not possible at many of the clinical centres. NAIVE USERS VIEW OF NETWORKING - Michel Jorda. Cystic Fibrosis Researchers in France have set up a mailing list on EARN (MUCO-FR@FRMOP11) and also connect via the French Transpac to FRSUN12 on EARN using Videotex terminals provided by the French PTT. Michel described their experiences and made several suggestions for improvement. 1. Keywords should be assigned to email distribution lists. It is not easy to tell the topic for discussion on a list from the list name. The command LIST GLOBAL gives a list of all email lists on all LISTSERVs - but without keyword searching capability it is extremely time consuming for a user to find those lists in which he might be interested. 2. EARN documentation should be written without reference to specific operating systems, terminals etc. to avoid confusing users. 3. Automatic addition of a list subscriber to a User Directory, using the keywords of the list as the "interest" keywords for the user. This would be with the user's permission of course. I pointed out that many of the suggestions Michel was making (better support for special interest groups, better user guides, the need for user friendly interfaces, keywords on distribution lists to allow sorting according to topic, etc.) had been brought up at EARN 88. The problem was not that these issues were not felt to be important, but that there was no mechanism for making progress on them, as EARN is a collaborative network. (But see note on Closing Session.) A3: NETWORKING IN BIOMEDICAL RESEARCH - V. Guardabasso. The problems of using the network encountered by special interest groups with little expertise in networking or computing was described. The same themes came up again: the need for better user interfaces, the need for access to networks via the public data networks to enable small clinics to connect cost-effectively, etc. Jean-Claude Salomon, who was chairing this session, reiterated the problems of drawing researchers into using networks and the facilities they provide for group communication. I pointed out that we (but again who!) need to go out and "spread the word". We need to start giving demonstrations at Medical (and other) conferences of the facilities available to researchers, with examples of how EARN (and other networks) are already being used by some groups to enhance the quality of their international collaborative research. B3: ASTRA - Stefano Trumpy. ASTRA is a joint project between IBM and CNR (Italian national research council). There was a presentation on ASTRA at EARN 88. Since then ASTRA has grown in functionality and scope and is running a production service. The main database STAR is a database of all the joint IBM-CNR research projects. Information is held about institutions, researchers, publications, software, courses etc. ASTRA stands for "Application Software and Technology Reports for Academia". It provides a distribution service for public domain software and project reports related to teaching and research activities. EARN users of ASTRA can get a local ASTRA user interface package. This enables the user to search several different geographically separated databases (via ASTRA) using one set of search commands. There are 20 databases currently accessible via ASTRA (mainly Italian ones) and future plans include access to ISAAC dbase in Washington, Soft-Info dbase (Univ Bridgeport), WISC-WARE (Univ. Wisconsin) and the NISS Software Catalogue (UK). Final Session ------------- The closing session was devoted to a panel discussion. During this it was announced that the EARN Technical Group were to suggest to the EARN Board that there should be an EARN Information and User Support Group. Hopefully this will be agreed and can be realised without a significant increase in the EARN budget. Personally, I would very much like to see this group form contacts with the other groups working in the area of Information Services: European RARE WG3 subgroup1 (information services and pilot European Information Service) and the subgroup of the US IETF Working Group on User Support. EARN 89 was generally very well organised and seemed to run smoothly. A couple of sessions were cancelled (one on Network Management) or rearranged, but I guess that was out of the hands of the organisers. The time-tabling of the sessions left plenty of time for talking to others in related fields, which is one of the main advantages of such conferences. I had some very interesting discussions with various people involved with information services: from EARN, NSFnet, SURFnet, HEANET, ASTRA, LIFESCI etc. We all welcome the increasing co-operation between user support staff on the various networks. Jill Foster 6th June 1989 ---------------------------------------------------------------------- Jill Foster Tel: +44 91 222 8250 Computing Laboratory Telex: 53654 University of Newcastle upon Tyne Fax: +44 91 261 1182 UK Email: Jill.Foster@Newcastle.ac.uk