Paul Bryant's Networking Correspondence
Dear Francisco,
Thank you for your letter. I am most anxious to help your work and to help you study our experience to select the best and avoid our mistakes.
I am dedicated to developing a Europe wide academic network to support our scientists at a price they can afford. My own belief, which I share with many of my European colleagues, is that we must concentrate on developing a network based on ISO protocols to avoid any suspicion of supporting any particular manufacturer or other parochial group such as my own UK Coloured Book community. I find such a project exciting.
You mention Ireland. In fact Ireland has now decided to adopt the UK Coloured Books and to take part in the migration of the UK to ISO standards. This is not an indication that I would advocate you taking a similar strategy unless it makes sense in your particular situation. Ireland has considerable historic connections with the UK which has led to quite a bit of network traffic and this has encouraged the idea.
You may be surprised that I am on the Board of Directors of the European Academic Research Network (EARN) which, if you've not heard of it, is supported by IBM and uses RSCS protocols. In fact Spain has links from the Politecnica in Madrid and from somewhere in Barcelona. CEPT has demanded that this network should migrate to ISO protocols. I am the technical director of EARN and have to take an active part in this migration which I fully support. My contact at UNIVERSIDAD POLITECNICA DE MADRID is M L Mate who may be of use to you.
Now, turning to your letter. I have given some thought to a possible set of people for you to see and the areas you may be interested in. These are:-
1. Rutherford Laboratory. I, as head of networking, will be delighted to show you what we are doing and how our thinking is progressing.
I would also like you to see Peter Linnington who is head of the university's Joint Network Team and looks after the academic network JANET. I would also like you to meet Dr. Brian Davies who is the director of computing for the Science and Engineering Research Council and could be of use to you. Peter can talk to you about tariffs. I am also interested in tariffs and Peter and my views are different! Since we have a gateway to PSS and EARN at Rutherford we can cover that topic.
2. Department of Trade and Industry. I would like you to see Keith Bartlett. He has been doing a lot of good work promoting the governments 'Intercept Strategy' which is aimed at encouraging industry to adopt the ISO protocols. I am sure your PTT colleges would value a meeting with him. His address is-
3. University of London Computing Centre. Roland Rosner looks after the networking in the London area and has a particular interest in IBM networking. He can tell you of the networking problems of a large centre and how they are tacking them. The centre is used by most London universities. He can talk to you about his own centre and how the various London colleges organize their local networks. His address is-
4. University College London- department of Computer Science. Peter Kirstein operates the gateway to ARPA and has several mail experts working for him including Steve Kille. This may be an interest to you and is easy to get at if you go to London to see the other two places. The address is-
5. Edinburgh Regional Computing Centre. This is a long way north but they are active in developing communications code. It may not be worth the long journey. They are principally interested in local area networks and gateways to them. Here again you could study how a large centre is run. The director Tommy Thomas (Dr. G R Thomas) could no doubt suggest who to see. Address-
6. Salford University. John Larmouth is a protocol expert and has a particular interest in Job Transfer and Manipulation Protocol. He has written a portable version of it which has had a slow birth. He is also running a name registration scheme on behalf of the community which is very interesting. The address
7. Bath University. John Thomas is setting up and Ethernet local area network and is knowledgeable on the problems on this and other types of network. This is a smaller but lively centre which would make a nice contrast to the bigger ones. Address
8. CAMTEC. You might be interested in visiting CAMTEC Electronics in Leicester. They produce out X25 PADs and have several other interesting products in particular small X25 switches and gateways to LANs. The man who runs it is Dave Dury.
I would not suggest you visit all these sites but the ones in the south of England could make a nice trip with little travel. Of course, you might like to visit the Heather Moors for a spot of golf or Grouse shooting so don't let me stop you!
May I also encourage you to visit CERN and talk to Francoir Flukiger of DD Division. He is their X25 expert and has been recently sorting out the future networking strategy on the site. He has a number of experts who can give you a view on the problems of a large international site. This would be of particular value as Spain is now a member of CERN. While on the subject, I am disappointing that Spain has not sent a representative to the ECFA (European Committee For Future Accelerators) subgroup 5 on links and networks. This is a lively and important group who have been trying to sort out networking on a European is not global basis for the high energy physicists. I comment the group to you.
Another important person to visit is Professor Zandar from the Hahn Mitner Insitute in Berlin. He initiated the recent 'harmonization' initiative which has been taken up by ESPRIT. The German situation is most interesting as they have 60 machines on EARN while at the same time they are setting up DFN and they thus have two academic networks which must be a worry.
Stockholm University is a good site to visit to get a view of what is happening in the Nordic countries and I am sure that Birgitta Carlson of QZ whould welcome you.
I would be most happy to set up the visits to the UK sites for you if you would indicate the sites you want to visit and the dates. It would be easier for me as I know all the people and I am afraid that recent experience with the Spanish telephone service means you might have difficulty.
May I also bring to your attention the European Networkshop. This will be in Luxembourg from 13 to 15. The conference is by invitation only but each country has a representative who will invite leading experts from the country. Your countries representative is G Thomas from ?. I suggest you contact him for details and I think you would find it of value and we would value your participation. We will be dedicating the conference to trying to forward the idea of an integrated Europe wide academic network based on ISO standards. I believe this will be a vital meeting and I am most anxious that the work will be supported by Spain.
I have pleasure in enclosing a form for the UK networkshop. It is a bit late to apply but I am sure we could find you a place. It is informal and cheap and you would meet a lot of interesting people. It might be that you could combine it with the trip to the UK. You would have to visit most places before Networkshop as Easter follows it and many people will be on holiday directly after it.
I very much look forward to renewing our acquaintance.
With best wishes,
Paul Bryant.
The meeting was characterized by less IBM participation and very lively discussion. The long thin meeting room made life difficult. As usual attendance from each country was excellent.
There are now 600 machines on the combined EARN / BITNET network. There are now plans for African, Near East, Far East and Australian branches. Ira Fuchs is intent on making it a real world wide scholarly network. No one liked to ask what the attitude to Russia joining would be although there was much discussion on the problems of having Israel and Saudi Arabia as members of the same network.
I, as usual, came in for considerable flack for the UK still having no permission to connect. However, I shall be the only member having a genuine written license. I shall be seeking funds to have the document illuminated and placed in a suitable casket!
The association is now a legal body incorporated in France and has had a donation of $50,000 from IBM. The Board considered how it should raise funds both from other manufacturers, governments and by memberships fees. I made the point that if there were fees imposed on ether users or sites then EARN will be in competition for funds with, in the UKs case, JANET and there may be strong pressure for JANET to win. An additional problem is how such fees can be levies and access controlled in the case of a network like JANET. None the less, the Association does need funds to support work done by members on behalf of the association. It seems certain sites are worried over the expense of being central EARN sites. Certainly Rutherford, or at least I on behalf of Rutherford, am worried over the expense and effort involved. (Group leaders please note).
This was considered a success. The Boards main discussion was on the migration to ISO. They invited me to lead a small group of 4 or 5 ISO experts from EARN to meet with IBM experts at La Gaude to discuss and plan the migration in accordance with the useful and encouraging letter from IBM. I intend and was encouraged to organize a weeks meeting with IBM to get the planning over in one concentrated effort. Again, GLM please note effort involved and expense. However the activity will be of use to JANET. I was also asked to talk to other manufacturers, possible the ESPRIT big 12 to invite other participation and coordination. I pointed out that the Zandar harmonization and Euro Networkshop must be taken into account.
Naming, as in all networks, had an overlong discussion and was as fruitless as most other discussions on the subject. Ira insisted, rightly, that there should be one set of route tables which will be held at CUNY. Otherwise the network address wise will become a shambles. It will be important for all RSCS route tables to come from CUNY and not to be changed by the sites. Interesting since the Saudi Arabian tables will contain nodes in Israel and visa versa! An interesting problem which engendered a long discussion as to what would happen if one or other objected. Ira felt that the network must not mirror political situations but others felt that some more pragmatic approach may be needed if EARN is not to engender a third world war.
It was agreed to invite cross representation between BITNET and EARN committees. Thus I will be invited to the BITNET technical committee as I am the technical coordinator for EARN. GLM again note that this implies effort and funds that should but probably will not come from the association.
A number of interesting developments are taking place.
CUNY is looking at conferencing systems in particular GRAND and POTACOM. The evaluation is not complete.
CUNY has a project to put RSCS over X25 which shows a lot of promise and may be the way EARN should migrate to ISO.
Ira told us that Science Net was to use TCP/IP and he believed that this was a step towards ISO. I find it a bit hard to follow that argument.
BITNET now as a CHAT system which allows a real time conference between people on terminals. It looks to be very popular and useful.
There is some concern over the use of CP/QU which allows one to find out who is on other machines and encourages people to go on 'fishing expeditions'. There are other 'fishing' techniques which are expensive in network resources. Note Doug- the best hackers have yet to arrive!
A fascinating discussion in which I lost popularity.
Hargan Hultch (Germany) felt that EARN should become the 'manager' of a European Academic network. If this was not the case and EARN would migrate and vanish as ISO came about then we might as well do as little as possible as we were running an ephemeral network. The view had a lot of sympathy.
I took the unpopular view that EARN did not have the authority to take on such a role and although it could offer its services in such a role it certainly should not be surprised if the offer was not accepted. i took the view that any European Academic network must be set up by the various academic authorities and have the appropriate representation from countries. On the other hand EARN was largely self appointed. There was a view that EARN Board of Directors was not self appointed in that at least Germany and a few other countries had had an election for their Board member by the participants. I cannot see that an election of a Board member by those who had joined EARN would be valid as representative of the Academic community.
The Euro Networkshop was discussed and I outlined its objectives. Three was a request for EARN to be invited. I questioned why EARN had any special position as opposed to JANET, DFN or any other network and that inviting EARN would encourage requests from many other organizations and it was the organizing committee's decision to organize it on a country basis.
It was clear that about half the meeting were disappointed that EARN was not progressing to be the European network, The other half were welcoming to Euro networkshop and to ISO migration.
As I suspected, now that the network is becoming established the Board has fractioned into the RSCS and the ISO supporters as people start to appreciate what the arguments are all about. The future of the group will be fascinating.
This will coincide with the Landwebber mail meeting in Stockholm in August.
As a result of the request for information of current local network activities and future requirements 4 replies were received from:-
R E Thomas Informatics Division C Reason Laser Division P McPherson S & A Division M Hapgood Geophysics & Radio
Details of Computing Division activities have been included by the author.
It is no surprise that the currently provided X25, PACX, and Coaxial 3270 service provides the majority of the services. Within the limitations of the technology these services appear to provide a reasonable spread of services which are reliable.
In several corners of the site there are various other private networks providing a variety of services. There is some commonalty in the services they provide but there is also a wide variety of protocols and technologies.
Surprisingly, there was little support for a site wide high speed local area network. Most of the demands were for modest increases in speed and for a wider variety of services such as the connection of microprocessors and a service to TELEX.
This is provided site wide. A policy decision some time ago was taken to connect all suitable computers to this network and to provide terminal access at up to 4800 BPS. This policy has been carried out and these facilities are fairly freely available.
The policy has favoured the extension of terminal services by the use of PADs rather than PACX connections. This has not been a vigorously applied policy as there have been some worries that the capacity of the X25 network may be exceeded.
The replies show that the majority of terminal access is via PADs. It is also clear that each division has a requirement to access a variety of equipment both on and off site.
The network is based on GEC switches. Consideration is being given by Computing Division as to how it should be ungraded. However no major changes are contemplated for a long time.
Computing Division installed a large Cambridge Ring which visited every office in the Division. It was expected that this would provide the basis of a site wide development. In the event, the non appearance of equipment and the escalation of costs prevented the development reaching maturity. The Ring now connects to all the GECs in the Division and a number of Diamond word processors. There are two Ring PADs and an X25 gateway, all from CAMTEC. The equipment, even though some of it is developed in house is reliable although it does not get much use. Over the last year or so there does not seen to have been much increase in interest in Rings. This together with the lack of equipment, small number of suppliers and uncertainty over its status as a standard has led the division not to consider its development for the time being. It uses CR82 protocols.
Informatics Division share the Computing Divisions Ring and it supports some PERQs, RLVD and RLDA. These run CR82 protocols.
SNS have a Cambridge Ring but no details have been supplied.
Computing Division have an experimental ethernet which is shared with Informatics Division. This has been used mainly by Informatics for RLVF, and SUNs and uses TCP/IP. Computing Division has been attempting to gain expertise in the installation and trouble shooting on Ethernets. The network is about to be used for developing a connection between the HEP VAX and the IBM. When complete a new cable is to laid between R1 and R27 if the current experiments prove the network to be capable of the performance required.
Computing Division also have an experimental Ethernet between two IBM PCs using Cheapernet which differs from the other version in using cheaper cable and components. So far only low level ISO protocols have been tried. This is being done with the knowledge that a demand for good network services between PCs and the IBM are likely and there is a desire to use ISO protocols.
HEP have an Ethernet connecting their VAX to the 370/E devices. No details have been provided. HEP are also believed to have a DECNET connection between some of their machines.
This service is provided on PRIMEs and GECs for file transfer between micros and mainframes. It has received a small but steady trickle of customers. Its non adoption by other sites makes it unattractive and the service is likely to die a natural death as superior services are provided. Only Informatics group claim to use it between PERQs and VAXes.
There is an extensive network of coaxial cables providing 3270 services to the IBM. All the replies apart from G & R mentioned the need for PROFs access. Curiously no comments were received for the use of 3270 by scientists. It should also be remarked that Full screen CIFER access was not mentioned.
For historic reasons the site is covered by a miscellany of cables. There is certainly a cable problem in that many buildings are not equipped with ducting and quite often ducting is jam-full. >
Computing division is putting a bit of thought into how the site should be cabled. A fibre optic backbone network is being looked at as well the the newly announced IBM cable system. Unfortunately very little effort is available for this work.
Although the request for information was not aimed at this area there were a large number of comments which warrant a separate section.
S & A Division find that the response via PADs can be slow during peak periods.
S & A see a small requirement for a 9.6K PAD service but G & R division find that even 9.6K is a bit slow for graphics and full screen terminals.
Laser Division would like to see their PCs on some sort of high speed network.
G & R would like to see the number of PAD connections increased in the Division to one per member of staff which is a modest request.
G & R and S & A saw a need for micros and word processors to be networked better. There was a particular request for the BBC micro and also the provision of KERMIT in particular for file transfers to the USA.
G & R and Computing Division are interested in TELEX to on site computers. This service would be of particular value out of hours.
A general comment was that Divisions wanted more of the same. Better access to overseas networks, a bit more speed, a bit more flexibility. The impression gained that the current direction of development was right.
Informatics and Computing would like to see a site wide mail service based on the pilot one operating in Computing Division.
S & A required to transfer 50 to 100 MBytes per day. It was not clear whether this was a wide or local area requirement. There is a possibility that they may get another VAX in which case they would probably connect then using Ethernet
Informatics Division have very extensive plans and full details are in the attached paper. There requirement is for a mixture of service and development. To this end they require Cambridge Rings and Ethernets for many of their machines. They intend to run a variety of protocols. some of these are dictated by the pragmatics of availability and others by a desire to follow standards.
Computing Division is only undertaking small investigations on the grounds that there is insufficient equipment available to mount any sort of service. There is a feeling that Ethernets look to be the best bet which is why some expertise is being sought. The feeling is that the current site wide services are only capable of modest improvements and that if a far higher quality is needed then a new technology is required. It is also seen that the current facilities are becoming a rag bag of techniques and it would be desirable to rationalize the situation. However this can only be done if the cost is reasonable. Any network must provide all the current services, including 3270, as well as being able to carry desirable services such as distributed computing ones.
G & R would need some convincing that a high speed LAN was justified particularly from the financial point of view.
From the replies it is difficult to see a general need for a high speed LAN. None the less there are several examples where high speed connections are needed but these tend to be special cases. These special cases are being met to a modest degree of satisfaction by special equipment.
The replies also indicate that modest improvements in speeds would be most welcome.
The conclusion would seem to be that if a high speed LAN could be provided offering, in general, modest performance increases and in a small number of cases big speed increases at a cost not greatly in excess of the current costs then it would be welcome. One can confidently predict that this will come about and one can say with even more confidence that it is currently impossible.
A question left to answer is what level of resources should be put into attempting to bring forward the installation of a LAN and what form such an activity should take.
Present: B Davies (Chairman) S Rows Laser Division P Bryant (Secretary) J Barlow HEP M W Johnson Neutron Division B Jones Scientific Admin A D Bryden Technology Division P McPherson S & A Division R E Thomas Information Technology J Harrison Geophysics & Rad P Linington JNT Apologies: J Hutton HEP M Hopgood Geophysics & Rad
B Davies stated that the committee had been set up by Division Heads Committee following an open meeting (P3) held in May 1984. This meeting would have a representative from each division.
A separate committee was being set up to coordinate the network activities of interest to all SERC sites which was a successor to the SERC Network Management Committee which had been disbanded shortly after the inauguration of JANET. The recent decision to start this committee made it less necessary for RCCC to consider wide area matters.
J Barlow thought that RCCC should not have any responsibilities for private networks in divisions and it should not be necessary to get any authorization from RCCC to undertake such projects. Dr. Davies said that at this stage it was important for the Committee to know what networks existed and what developments were in progress rather than to become preoccupied with authorizations or approvals.
After discussion it was decided to replace the first two sections of the terms of reference (P1) by the essence of the first five sections of (P2) from P Linington. A new document will be produced.
Action: P Bryant
It was decided that information was required on existing networks and projects. This will be collected by members providing details of activities in their divisions to P Bryant who would produce a summary document. The information required includes:-
Details of existing networks and its equipment Indications of performance being obtained Comments on how it meets requirements Problems in operating the network Who is responsible for the network Comments on reliability.
, P Bryant
A further document will be produced on the future requirements as seen by members. This should be based on the requirements for two years time and five years time. Members should send contribution to P Bryant who will produce a summary document. The information required includes:
New function required Traffic levels Interconnection requirements.
Action: All , P Bryant
Monday 29 April 1985 at 14.00 in Think Room R26 (Note revised date)
It was agreed that the meeting would consider the traffic requirements on the wide area network arising from activities on the RAL site.
ACTIONS MEETING 1 M1.1.1 Produce new terms of reference P Bryant M1.2.1 Send details of existing networks to P Bryant All M1.2.2 Produce summary of existing networks P Bryant M1.2.3 Send details of network requirements to P Bryant All M1.2.4 Produce summary of network requirements P Bryant
Dear Alain,
The recent EARN Board of Directors' meeting considered the letter from Dr. Herb Budd concerning the migration of EARN to use ISO protocols.
The meeting welcomed the suggestion that the Board of Directors set up a committee to define how the migration should be under taken and were pleased with the kind offer of IBM to give us the assistance of experts from La Gaude. As a consequence I have been asked to arrange a meeting or series of meetings with a view to exploring the many factors affecting the issue and recommending how the migration should be under taken.
I think it might be more profitable if we arrange a relatively long meeting with a small number of people and try to produce a preliminary proposal then and there rather than have a series of meetings which will prolong the process. I do not expect that we can finish the job at such a meeting but it may well be that the remainder of the work can be done by mail (preferably electronic) as hopefully we should have set up a good working relationship. I am expecting that 4 or 5 representatives from the academic side will be able to take part with perhaps a couple of IBM experts.
I am thinking of a three day meeting. Since good access to expertise and documentation is essential, it would be highly convenient to have the meeting at La Gaude. I would like the meeting as soon as possible so that a paper can be put to the Board of Directors meeting in July.
I shall be preparing two papers for the meeting. The first will be an updated version of the Migration to ISO paper, which has already been circulated to the Technical Group, and the other will be a statement of requirements which I am currently working on. I think that the meeting will have to start with a few short presentations of the possibilities and the requirements.
The possibility of this meeting taking place may hinge on whether the Association is willing to fund it. As I expect you have seen, the sites involved with EARN are often stretched with respect to money and manpower and this may effect whether we can get the right experts to attend.
At this stage I would value your comments on the proposal and if it acceptable to you can you suggest who I should liaise with in IBM.
I felt that the recent BOD meeting was a great success and we are now working well together and tackling the many difficult problems that face us.
For your interest I enclose a copy of 'Network News' which has a wide circulation around the universities and which contains an article on EARN.
With best wishes
Paul Bryant. EARN Technical Coordinator.
27 March. Arrive Heathrow IB342 at 18.15. Transport to Crown and Thistle- Abingdon. 28 March. Visit Rutherford Lab. P Bryant, P Girard, B Davies and R Cooper. 8.30 Car from Crown and Thistle to Rutherford Laboratory. 9.00 Political structure of computing and networking in the UK academic community. 9.30 Current structure of JANET. 10.00 Coffee 10.15 Demonstration of Janet followed by tour of equipment. 11.00 Network management and operation of the network. 12.00 Lunch 13.00 Implementation problems and IBM network code. 14.00 Relations with other networks- EARN, CERN and ARPA. 15.00 Work of the JNT. 16.00 Transition to ISO protocols. 16.15 Possible Anglo Spanish cooperation. 17.00 Return to Crown and Thistle 29 March 08.00 Leave for Didcot station. Catch 08.54 for London arrive Paddington 09.37 catch Circle line Underground for Victoria (40p) 10.00 Visit to Department of Trade and Industry- Brian Wood 01-213-5464 29 Bressenden Place. Very near Underground station. 12.00 Leave for University College London- Gower Street. (Euston Square Underground) 12.30 Lunch with Tom Daniel. (ask for Computing Science Department, Department Office) 01-387-7050 15.30 Leave For ULCC. Walk or Russel square Underground. 20 Guilford Street. 16.00 Visit University of London Computer Centre- Roland Rosner. 01-405- 8400. ??.?? Return Abingdon from Paddington. 17.30, 18.40, 19.03, 19.50, 20.35 21.05, 22.05, 23.05 or 23.40. 30 March free suggest look round Oxford and catch train for Leicester from Oxford. You will have to change at Birmingham. Plenty of trains but slow. Booked in At International Hotel within walking distance of station. 9.00 Visit CAMTEC. You will be collected by Mike Stevens. There will be a presentation of a range of PADs, switches and other network components. CAMTEC will provide transport to Networkshop at Sheffield.
An exciting meeting! As usual BT came out punch drunk.
TELEX yet again came up. We learned that there are 73 users of the TNA and that if they had 60% more it would be profitable!!!!!!!!! Our comments were that with the technical spec of the TNA and the penal tariff we were surprised their were any customers at all and that if the technical problems and the tariff were sorted out there would be thousands of users. BT still seem fairly obdurate. One comment was that they did not want the TNA to undermine their TELETEX service to which the answer is - what TELETEX service! Apart from the satisfaction of scoring points from BT the argument got nowhere.
CTIG was worried over the proliferation of PADs from BT with different protocols. It was worrying that some of these did not follow international protocols. Block mode working came up for special study and we learned that BT with EPAD and the Germans with a protocol geared to the use of videotex were attempting a CCITT standard.
The Transition meeting asked us for comments on the transition to triple X 84. We came to the following conclusions:-
We were told that it was now possible for anyone to go to CCITT meetings. CTIG welcomed this and determined to send a representative to CCITT working meetings when relevant topics were being discussed. JNT may be asked to support such attendance. We are not 100% sure that our information was correct but we will soon find out!
Regretfully the attendance was small at 7 with only one industrial representative.
Entry Assist is a set of facilities recently provided for IBM 3270 terminals. The facilities were not known about when the 'full screen CIFER' was defined. This paper defines the extra facilities for the CIFER and the IBM PC.
Entry assist is defined in IBM document 'GA23-0119-1'. It provides:-
The facilities provided by entry assist fall into two classes:-
In the full screen CIFER the first class of facilities are permanently active.
As the keyboards of the 3270, CIFER and IBM PC are very different each device will implement entry assist in accordance with the 'style' of the device.
Facility CIFER IBM PC Tab back | |<- * Word tab forward Ctrl -> * Word tab backwards Ctrl <- * Word delete | Del Toggle entry assist Alt PF6 Toggle word wrap Alt PF5 Toggle set up Alt PF4 Set left margin < Set right margin > Set audible end of line * Toggle alternate right margin End Set tab |<-_ Clear tabs Clear Toggle change scale line Home
The cursor position will be permanently displayed.
The state of the entry assist toggle and word wrap toggle will be displayed on the 25 line.
Apart from statements above the action of entry assist will be as defined in the IBM document.
Note that margins and tabs will not be changed when going from scroll mode to full screen mode.
A condition under which EARN is to be allowed to operate is that there will be an evolution towards the use of international ISO and CCITT protocol standards. In addition CEPT would prefer EARN to operate over public data networks.
If EARN is to use public switched services, ether X21 or X25, then CEPT and the PTTs can legally only require that protocols demanded by the particular technology are used. Thus, in the case of the public packet switched network only X25 is needed. Other CCITT or ISO protocols, such as X400, need only be provided if EARN decides to base services on them. There is no legal reason why any private or proprietary protocols should not be used over the public networks.
If EARN is allowed to use leased lines then it is technically possible to use proprietary protocols such as RSCS or non proprietary ones such as X25. It is also possible to use a mixture of protocols such as RSCS over X25. Of course, the use of X25 would need packet switches to be installed and the use of RSCS over X25 would need some adapting of the RSCS protocol. There is advantage in following a strategy that allows EARN to use either public or private networks in that it allows the network a wider range of choices to achieve performance and economic aims.
However CEPT could demand that EARN, unlike other subscribers, did follow certain high level protocols permanently as a condition of allowing the current use of leased lines. It is difficult to see how such conditions could be imposed as, if and when EARN has migrated to use public networks, it would be difficult to identify EARN as a particular set of machines.
The above analysis suggests that there are three areas which should be differentiated:-
The CCITT (1) and ISO (2) protocols are designed to provide a rich set of facilities to meet the needs of a wide range of uses. This has led them to be fairly complex and also to have a number of options. To achieve successful communication the subscribers have to provide compatible realizations of the protocols at each end of a connection. In practice this means providing a common subset of options. Also any intervening entities, such as switches and gateways, must also be compatible.
The protocols are designed to be open ended in that new protocols may be added as needed although the ones already defined are expected to be very stable. Currently the X25 (version 80) is widely available both on computers and on public networks. The X3, X28,and X29 protocols over X25 are also widely available. Although ISO has been attempting to define other protocols it is only in the last year or two that a sufficiency of protocols have been available for other applications. The X400 or Message Handling System protocols are now complete and a number of implementations are available. The file transfer protocol, FTAM, which can also be used for file access has reached its second draft as a 'Draft Protocol' and is now stable. No implementations are yet complete although one or two can be expected by the end of the year. Virtual Terminal Protocol is still some way off and this protocol is important is high grade terminal facilities are to be provided like 3270 or VT100 facilities. Although FTAM can provide job transfer facilities the Job Transfer protocol gives far more sophisticated facilities to allow jobs to be traced and controlled in various ways. TELETEX is a well defined protocol which is aimed at one day replacing TELEX. It is very closely related to X400.
A rather worrying problem is that X25 has been extended in the 1984 version and the PTTs have plans to change to this version in a few years. The protocol has one major advantage in that it includes extra addressing mechanisms which are important for getting through gateways. This is of particular importance in the academic community.
The advantage of using ISO protocols is that they will, hopefully, be available on a wide range of machines. There are two principle disadvantages. Firstly, their use may lead to some loss of quality of service on any particular range of machines, since the manufacturer is highly likely to put more effort into his proprietary network schemes which bring in revenue. Secondly, the implementations of ISO protocols, especially at the higher levels, are not yet widely available.
Several other organizations are interested in fostering services using the ISO standards and care must be taken to ensure that full account is taken of their activities. The organizations are (this is not an exhaustive list):-
There may well be similar groups in existence. Certainly there is an intention in the groups that there should be a single set of inter working recommendations. There is considerable cross representation between the groups.
The state of implementation of the protocols varies from protocol to protocol and machine to machine. It has also been found that in several cases lower level protocols have been provided which will only support specific and often proprietary higher levels. Thus it is currently difficult to provide working networks encompassing a wide variety of machines.
A cause for concern is that the various PTTs have adopted different 'flavours' of X25 which has meant that host implementations may only operate on a subset of networks or, rather more commonly, the provider gives a range of options to allow his product to be tailored to a specific network. There are pressures from several international groups for the PTTs to resolve these problems.
Implementations of triple X are a constant source of amazement in that their quality is very often pitiful. None the less, better quality products are becoming available and in general shortcomings in implementations can be overcome (6).
Above the network layer the implementations of ISO or CCITT protocols are few and far between. There are a few versions of the MHS (Message Handling Protocols also known as X400) but their quality seems good. There are implementations of the Teletex protocols although these are not of much use in shipping files or mail over networks like EARN. Most examples are on workstations geared to office use. There are a few implementations of transport layer and of session layer many of which are buried in the MHS and TELETEX products. Implementations of transport service and session layer are being produced by several manufacturers. An urgent requirement is the provision of a file transfer protocol so that an ISO network can provide comparable services to EARN.
In summary there are a sufficiency of ISO protocols now defined to provide the facilities now available in EARN. The difficulty is the state of implementation.
These options assume EARN must use public networks.
There are two possible types of public network- X21 and X25.
The use of X21 depends on the provision of international ISDN connections which are highly unlikely in the near future. In addition only a subset of European countries will have such networks in the near future. IBM's plans for providing services are unknown although it appears that recent hardware announcements, such as the 3725, make such connections possible.
Most manufacturers now offer some sort of X25 connection. In some cases other suppliers can provide implementations often of a portable nature. In a number of cases implementations have restrictions which limit their usefulness. The list of implementations below is not exhaustive and is probably not accurate.
IBM 37X5, Amdahl 4705, ITT and similar equipment. Although these products support X25 connections they can only be used within an SNA environment. SNA support varies with different systems and the recently announced SNA support is directed to interactive triple X services in that it has no interface to NJE. From the academic point of view the use of SNA has the problem that it consumes considerable computing and manpower resources. In addition SNA is not yet available under the popular VM/CMS. COMPRO provide modifications to the 3705 or 4705 code to give a means of providing a level 3 interface which may be used to support other protocols in a non SNA environment. In particular in a VM environment. IBM are committed to provide ISO transport and session protocols. These will be provided via 37X5 equipment. Unfortunately this is only announced initially for MVS systems (7). The products are likely to be part of the SNA products. Few details are yet available.
IBM Series 1. This can provide a level 3 service. It tends to be an inconvenient solution as IBM are unlikely to develop this product. It also means that each site will have to acquire a Series 1 which is expensive. IBM are supporting the development of 'coloured book' protocols for IBM computers using this product. This work is being done by Salford University and is currently being field tested at Liverpool University. Performance is not yet known but documentation is available (8).
Dialine. Dialine offers a package called CSX, which includes a special channel coupler and X25 interface from CSI installed on a Series 1. This emulates a 3705 running the EP with BSC lines. In this way RSCS can be, in effect, carried over X25 connections.
The 'Edinburgh Box'. Rutherford have had an X25 connection into their IBM computers for many years but this has used bi sync at link level. Edinburgh University have produced a cheap micro computer based 'box' to convert bi sync to HDLC. Thus a level 3 interface can be provided by this means. The equipment is only suitable for a low throughput of 9.6K and costs are between 2500 pounds and 4000 pounds. The solution has the disadvantage that it uses non manufacturer supported software.
VAX PSI. VAX computers have a good DEC supported X25 product and there are no problems in supporting higher level protocols. The company also provide a reasonable triple X product (9).
CDC. There is a CYBER NOS product which CDC provide in the UK. It is based on a 2551 front end. It is understood that versions of it will be available in other countries.
UNIX. Considerable resources have been going into producing X25 for UNIX with mixed results. Unfortunately early versions of UNIX have had a poor structure for introducing complex network products which has encouraged the use of front end processors with only moderate success. A newer version, Berkeley 5.3, is expected to have X25 embedded in it which will hopefully solves the problems.
There are many other computers offering good, bad and indifferent implementations which are currently of less interest to EARN. PRIME, GEC(UK), NORSK, Honeywell (Multics), ICL 2900, Harris (VOS) and PDP11 (RSX11) all have manufacturer supported products.
There appear to be four possible sets of protocols which can be run over X25. These are ISO, RSCS, SNA and Coloured Books. Mixtures of these may be possible in some cases.
IBM are committed to provide ISO transport and session layers. As yet the dates for these are unknown. The initial release will be for MVS with VTAM and JES2/3. No plans for other systems are known but as the products seem to be embedded in SNA it is likely that a VM product will be available when a VM version of SNA becomes available. Plans for MHS or Teletex as well as the higher level ISO protocols are unknown. The University of Waterloo (Canada) are believed to be working on an implementation of MHS to operate under VM. British Columbia have produced MHS for VAX VMS and VAX UNIX which are of a good quality. Plans to provide ISO protocols are being formed by other manufacturers but are not known to the author (10).
IBM's Heidleburg Scientific Centre GDM Darmstadt have Developed an implementation of the OSI levels 4 and 5 using the X25 support in a Series 1. They state that it would be easy to develop it to provide a version of RSCS over X25. IBM has indicated that it would be willing to supply this, but of course future support would be a problem. The attractions of such an approach are clearly that it provides a migration path which would have a minimal impact on the user. The disadvantage is that it is unlikely that other manufacturers will provide RSCS over X25. Since there will be VAX, CDC and possibly other non IBM equipment in EARN similar products would be required on these machines if they were to be encouraged to remain members of EARN.
IBM supports SNA over X25. Since SNA is a well supported and developing product, basing EARN on it is attractive. As with RSCS, implementations of SNA over X25 for VAX and CDC would be desirable and plans for such products are unknown but are quite likely to appear. Experts claim that SNA requires greater expertise to operate and also consumes more computer resources than RSCS protocols. It must be added that there is very little known about the comparative performance of protocols or of their implementations.
The UK Coloured Book approach could be adopted. One is reluctant to recommend this as the UK community is about to embark on a migration to OSI protocols which is expected to take several years and EARN would have to follow a similar migration. The advantage of the approach is that a large variety of machines could be catered for immediately. CUNY are developing a scheme for running RSCS over X25. The product is at an advanced stage of development but few details are to hand.
Leaving aside PTT or CEPT objections it would be possible, if and when EARN had migrated to X25, to operate over a private X25 network or the public ones. Indeed it would be possible to have a mixture. The factors to be taken into account in deciding are many and complex but really boil down to regulations, finance and management.
If a private network were selected then an X25 switch or switches would have to be bought and installed. Experience suggests that such a network would need central management to ensure the integrity and operability of the network. This would require a level of cooperation above that required for the current EARN network.
Over time the aspirations of users are likely to rise and higher speed lines and switched may well be required. There will therefore be a need for continual capital funding. In the case of the use of the public network one would expect such expansion of the network to be financed from tariffs. In the private case other sources of funding would have to be sought which experience suggests is time consuming and liable to failure. It should be added that the public packet switched networks, particularly on international links, have a poor performance but one hopes this is temporary.
The use of public networks removes all legal restraints on the traffic carried and on the sites communicated with. It is likely that any private network will, at least initially, be constrained as EARN is likely to be constrained. Whether CEPT will be able to offer more attractive terms is debatable but probably doubtful.
As a private network grew it would develop gateways to other networks. In fact it could well be that EARN could develop into a high level network whose only reason for existence was the interconnection of various national networks. How likely such a development is depends on how national networks develop. For example, countries which decide to base their academic networks on the public networks would see no reason for using such a high level network for accessing similar systems (ignoring tariff issues).
Experience in the UK suggests that running a private X25 network is a complex and difficult task which really requires dedicated management. EARN would do well to study the UK experience before following such a path.
It is now clear that the regulations demand that EARN be regarded as a 'special' network. Although regulations may well change it seems unlikely that such changes will allow EARN to exist without a license from each authority.
The regulatory authorities have the problem that if EARN is given a license then other organizations can request, if not demand, similar treatment and this could have a serious effect on their revenue. The only argument EARN has for special treatment is that the traffic is of a non commercial nature. The granting of licenses on this basis does mean that certain types of traffic, which it would be useful to permit, are barred. For example, a commercial company cooperating with a university in areas where profits may one day be possible.
The PTTs appear to be determined to retain flexibility on any tariffs imposed on EARN. EARN will therefore be vulnerable to tariff increases (or decreases) and will be in a weak position to demand tariffs related to published ones.
In deciding between a public and private network the development potential and the tariff issues must be fully explored.
This paper has hopefully covered the main options for migration. If this is not the case the author is anxious to amend the paper to take any further options into account. Only brief details of each option have been given and these may be incorrect. Again readers are invited to provide corrections. Further details of options can be found in the references.
The next stage is for experts to consider each option further with a view to recommending one to the EARN Board of Directors.
(1) CCITT produces recommendations for the PTTs on a four year cycle. Recommendations for 1980 for the X series protocols up to X29 are available in one document which can be obtained from:- International Telecommunications Union, Place de Nations, Ch1211, Geneva, Switzerland.
Recommendations for 1984, which include X400, and being prepared.
(2) ISO protocol documents are available from national standards organizations. Some are 'draft standards'. Draft standards are unlikely to change before becoming standards. CCITT and ISO protocols are expected to converge and in some cases have already done so.
(3) The five COS documents which have already been produced are:-
COS1- Network layer. COS2- Transport layer. COS3- Local area networks. COS4- Triple X. COS5- Session layer.
They can be obtained from:- Information Technologies Task Force, Commission of the European Communities, DG III/B/1, Rue de la Loi 200, B-1040 Bruxelles, Belgium.
(4) The 'Coloured Book' Documents are:-
Yellow Book- Transport service. Blue Book- File transfer protocol. Grey Book- Mail protocol based on RFC 822. Red Book- Job transfer and manipulation. Green Book- Recommendations on the use of X3, X28 and X29, Pale Green Book- Triple X implementors guide.
DTI also produce three implementation documents:-
Technical Guide TG100/1- OSI network layer. Technical Guide TG100/2- Provision of OSI network service over X25. Technical Guide TG100/3- OSI transport layer.
These documents are available from:- IT Standards Unit, Department of Trade and Industry, Kingsgate House, 66-74 Victoria Street, London SW1E 6SJ, UK.
(5) Message Handling System or X400 protocol has been produced by CCITT. It is designed for inter message box traffic in particular over public networks. It is similar to RFS 822.
(6) See the Pale Green Book and COS4.
(7) Product information for OSI 4/5 support- Document GH12-5150. Available from IBM.
(8) Details of the coloured book products provided with IBM support can be obtained from:- System Research Department, University of Salford, Salford, Manchester, M5 4WT, UK.
(9) Details of the DEC X25 product, PSI, and their Blue Book file transfer product are obtainable from the company.
(10) The following documents were obtained from the University of British Columbia:-
The EAD distributed message system- user manual. EAN; a distributed message system, (Newfed) A strategy far a national electronic messaging system for research in Canada, (Gilmore).
These documents were obtained from:- Distributed Systems Research Group, The University of British Columbia, 2075 Westbrook Mall, Vancouver, British Columbia, V6T 1W5, Canada.
Electronic mail, like the telephone, the TELEX, and the postal services, shows signs of becoming all pervasive.
As with these services the customer expects to receive electronic mail from anyone without any pre arrangements and free of charge. He expects to be able to send mail to anyone, also without pre arrangements but expecting to be charged. He wants (with marginal expectation) the user interface to mail systems to be the same for all systems to avoid re-learning on moving between machines. It is possible to make telephone calls and write letters from anywhere in the world with only minimal training in each country, he wants electronic mail to be similar.
In becoming 'all pervasive' the telephone, TELEX and posts have arrived at 'standards' which are internationally recognized. Ask the question- where would we be if we had several telephone companies with different standards so that they could not interconnect? No doubt every desk would have a battery of telephones or a call may have to go via gateways to sort out the numbering problems and electronic differences that may be present.
Thus, if electronic mail is to become all pervasive the standards issues will have to be solved so that the customer can address mail to anyone else without the current addressing gymnastics. It is also desirable to have a rich, adequate and uniform set of facilities. None the less, as with the telephone, it is reasonable for the affluent user to have more extensive facilities within the standards whereas the less well off may have more primitive facilities but again within the standard.
These comments on electronic mail are clearly 'motherhood' statements. The question to be addressed here is 'can they be achieved' and 'by when' and 'how' as well as 'what do we do in the mean time'.
Although 'models' and 'protocols' are boring topics some elementary grasp of the basics is essential for an understanding of the subject.
The model for the telephone, TELEX, posts, and electronic mail are essentially the same.
The model has three levels- at least for the purposes of this exposition- and these are:-
This is roughly analogous in the postal services to typing, reading, filing and throwing away mail. The User Agent communicates directly with the user to find out what he wants and with the Message Transfer Agent to get the mail sent.
A particular implementation of a mail system may provide the tasks outlined above to a range of qualities. For example a User Agent may be almost null leaving the user to construct mail and its headers with an editor. The Message Transfer Agent may be incapable of the relay function or may not provide distribution list functions. These failings make mail less convenient rather than impossible. The only functions actually demanded is the ability to construct a document with a legal header and send it to another Message Transfer Agent and the ability to accept mail from a Data Transfer Agent and put it somewhere the user can access it. In fact, some implementations of mail provide little more than these very primitive facilities and still prove useful.
Leaving aside mail systems that cannot work over networks the remainder fall into two classes. The first are the proprietary ones such as the one DEC provides with DECNET and the one IBM provide with PROFS, and the non proprietary ones such as RFC822 from ARPA and the one defined by CCITT called Message Handling Service or X400 which are designed to be common to many types of machine.
The proprietary systems tend to be very easy to use in that one does not have to bother with the problems of having to traverse several networks or to be able to deal with complex naming and addressing schemes. On the other hand they have great difficulty in inter working with other mail systems and tend to rely on proprietary network scheme. A familiar example of this problem is the difficulty of getting PROFS mail to inter work with Grey Book mail due mainly to the problems of constructing addresses. In addition, the proprietary mail system may well make use of network features peculiar to his own network scheme.
There are really only two non proprietary mail systems.
The first was defined by ARPA and is called RFC822. This system has also been adopted by JANET and by BITNET/EARN. All three networks use different Data Transfer Agents. The only problem this poses is that mail has to be staged or relayed through a relatively simple gateway when going from one network to another with a different transport agent. In the case of JANET the transport agent is the Blue Book file transfer protocol and in the BITNET/EARN case an RSCS file transfer. RFC822 is the definition of a Message Transfer Agent and does not, regrettably, define a User Agent and so each implementer has invented his own. This is not so surprising as defining an interface that would work on a line at a time terminal and a 3270 and a VT100 could be a problem. However, MIT has provided a de facto standard which has been adopted by BITNET/EARN, the GEC mail system, Edinburgh and some other systems and sites. The author is a strong supporter of this standard and is disappointed that it has not been adopted more widely.
The second system is the X400 standard which is an international one. The primary aim of the standard is for linking public mail boxes provided by the PTTs. However this does not prevent it being used more widely. In fact, it is felt in the JANET and other communities that the X400 will be universally adopted in the long term as the only mail standard just as there is only one standard for telephones. A full definition is now available for use over X25 and a number of preliminary implementations exist. Implementations also exist over DECNET, TCP/IP and TTXP as the Data Transfer Agent.
The strategy is currently to use Grey book mail. This is a good strategy in that it enables easy mail transfers to the many other sites in the UK running the system and also to EARN (when the license is obtained). The standard is sufficiently good and unambiguous to allow very high qualities of mail from the facilities point of view. Unfortunately the full standard together with good directory and relay facilities has rarely been implemented but this has not inhibited its use. The un- availability of a definition of a user agent has been a nuisance in that it has encouraged implementers going on an 'invention spree' rather than preventing the development of services. The GEC implementation tends to be one of the best and this is why the mail system in Computing Division is based on it. In fact many machines send mail to the GEC mail machine for relaying so that it can take advantage of its sophisticated facilities. The IBM system is rather poor which is not surprising considering the small amount of effort expended on it in comparison the GEC one. The IBM mail agent is only capable of delivering incoming mail to a mail box and will do no manipulations of the headers. On outgoing mail there is no mail agent and the user agent sends the mail directly to Blue Book FTP. Thus there are almost no sophisticated facilities. This is done with a XEDIT program. This is not a serious problem merely inconvenience as the mail can be sent to the GEC to use its sophisticated facilities. None the less a better system would be of great value.
The PRIME and VAX mail systems fall somewhere between the standards set by the GEC and IBM.
Thus, there exists on the site a quite sophisticated and almost adequate mail infrastructure for the computer users.
A serious problem is PROFS. PROFS has its own mail system. This was initially developed for use inside a single machine and is now extended to operate over RSCS links between a set of PROFS supporting machines. It is seen erroneously by some as a competitor to Grey Book mail. This is wrong as the system is not implemented on other machines and has difficulty operating over networks other than RSCS because the relevant headers cannot be constructed except by a rather tortuous method of using a 'nick name' facility which allows a PROFS destination to be transformed from a string with a maximum length of 18 characters to the semi infinite string needed for a general RFC822 destination. The problem is that it is highly inconvenient to have to add to the nick name file every time one wants to send or reply to mail from a new correspondent. In addition, easy forwarding and reply facilities are not available. There are also problems as to whether mail should be seen as a 'document' or a 'note' which is said to be important especially if the mail originates as a TELEX message.
Some study is being under taken to see what options there are for improving the PROFs access to Grey Book and it is not clear whether a fully satisfactory solution is possible without IBM cooperation.
A further and complimentary study is underway to see if the Message Transfer Agent from Columbia University and a User Agent from MIT could be used to improve the IBM mail service. These are in use at CERN and their adoption would be a useful addition to the convergence of the CERN and Rutherford VM/CMS systems.
Grey Book mail is available on most machines in the UK academic community which are networked. It is highly unlikely that any other network mail scheme, apart from X400, will be used in the community since any other system will have few other sites to communicate with and, moreover, will attract adverse comment from JNT. Allowing alternative systems goes against the UK academic policy to support open networking. It would divide the community into non inter-communicating sub groups which would eventually lead to a demand for expensive and inconvenient gateways to provide inter-communication.
In the medium term it is expected that Grey Book will be phased out in favor of X400 which is a true international standard.
It is important that Rutherford support Grey Book and the eventual migration to X400 for if it does not the scientists using our machines will be denied the benefits of electronic mail to other sites both in the UK and abroad.
The practical problem for Rutherford is PROFS. It is important that effort is put into PROFS to allow it to use Grey Book mail effectively. It is the authors view that PROFs must adapt to Grey Book and eventually to X400 as near PROFs as possible as Grey Book and X400 are the standards and it is desirable long term to prevent islands of alternative mail arrangements existing in the community. Such a statement must be coloured by reality in that there are arguments for preserving the PROFs user mail interface and it may be difficult to obtain the 'quality' of service required if user to user mail on a single PROFs machine were to use the Grey Book or X400 Mail Transfer Agent. This needs study.
It would also be useful to improve the IBM Grey Book mail service for all CMS users and the best way of doing this may be to adopt the Columbia Mailer which is in use at many EARN and BITNET sites and provides a fairly good service. In addition the MIT mail interface has been similarly adopted and it could provide an improved user interface here. Effort will be required to adapt the mailer to Grey Book mail which is slightly different to RFC822. This work is, in any case, needed for the EARN gateway and will ether be done at Columbia or funded by IBM.
It is important that Rutherford should take a strong interest in the migration to X400 and to this end a recently released version for the VAX is being looked at experimentally. It is hopped that an IBM version will be available from Waterloo University some time this year and this should be investigated.
There is a request for TELEX facilities to and from PROFS. If TELEX were provided directly into PROFS then it would not be easy to have an automatic TELEX service from other than PROFS users since PROFS does not have the relay function needed. On the other hand a TELEX service to and from Grey book could be used by PROFS. In fact TELEX needs to interface to a Message Transfer Agent so that mail from anywhere, including PROFS, can be relayed to TELEX and visa versa. The TELEX system should attempt to send TELEXes to the customer and only if this fails should it go to some mail box- which may be a PROFS one- for forwarding manually.
As the PTTs will be adopting X400 there are good expectations that they will provide X400 to TELEX to TELETEX gateways which will remove the need to have TELEX or TELETEX machines. When these will appear is unclear. The PTTs will be using X400 to connect the various public mail systems such as Telecom Gold. This connection could also prove useful.
CERN is putting considerable effort into migrating to X400 and a report, called 'COMICS' has recently been produced which is well worth reading.
The international Academic networking group had its second meeting in Paris last summer and it was evident that experts on both side of the Atlantic were expecting X400 to solve the many inter working problems which exist in the form of inconvenient gateways. Correspondence on Swedish KOM reinforces this view as the resulting discussion had been almost exclusively about the problems of adopting X400.
202 Greece 204 Netherlands 206 Belgium 2062221006 PDP11, Paul Van Binst 208 France 208101 Transpac Packet gererator 208091000309*CISIFMST CISI PAD@1200 TO IBM 208091000270*CISIFMST CISI NOS 208091000519*CISIFMST CISI PAD@1200 TO IBM 208091010320 CIRCE NOS 208091040047 Bryant lpche 212 Monaco 2141 Spain (not quite yet) 222 Italy 2223077*D EURONET 2223078*D Euronet 2222620021*D Euronet 2222620022*D Euronet 228 Switzerland 2283101 Euronet 2284681140510 CERN LO=300 ME=1200 HI=4800 232 Austria 2341 UK (IPSS) 2342 UK (PSS) 234212300120 Dialnet leased line to states 23421230012011 Dialine 2 (type k to connect) 234219200001 Network Monitoring Centre (NFS) 234219200002 Network Monitoring Centre (NFS) 234219200190 Pergamon - Infoline (NFS) 234219200203 IP-SHARP (NFS) 234219200222 British Library Information System 234219200300 University College London (NFS) 234219200390 Lloyds registry IBM TSO (LOGON TCZMAB) 23421920100530 BT mysterious number (NFS) 23421920101013 Hostess system (BT) (NFS) 23421920101030 British Telecom (NFS) 234219210050 BT Mailbox facility (NFS) 234219511311 GEC Computers Ltd. 234219709111 National Physical Laboratory 234219806160 Queen Mary college London (NFS) 234220641141 University of Essex 234221222122 MIDNET Gateway at Birmingham 234221222223 Prime R & D at Birmingham (NFS) 234212300120 DIALNET leased line to States 234222236236 University of Wales 234222339399 University of Cambridge 234222530303 South-West Universities 234222715151 University of Kent 234223440144 Prime R & D at Bedford (NFS) 234223440345 Texas Instruments Ltd 234223519111 Atomic Energy Research Establishment at Harwell (NFS) 23422351911198 ADA UK Database 234223519191 Gateway to SRCNET at Rutherford 234227230230 University of Bristol (?) 234227230333 Avon Universities Computer Centre 234231354354 Edinburgh Regional Computer Centre 23423135435422 Edinburgh Regional Computer Centre 234233458158 University of St. Andrews 234239232323 University of Exeter 23423923232304 University of Exeter 234241260106 Strathclyde 11/40 RSX system 234241260260 University of Glasgow (NFS) 234246240240 ICL at Letchworth (NFS) 234251248248 University of Liverpool 234253265165 University of Leeds (NFS) 234260227227 MIDNET Gateway at Nottingham 234261600133*DLOCALLDF Salford IBM 4300 full screen Cifer 234261643143 University of Manchester Regional Computer Centre 234261643210 Salford computer 234261643343 Ferranti Computer Systems 234263259159 University of Newcastle 234270712217 Hatfield Polytechnic 234273417171 DEC at Reading 234273417217 Modcomp 234275317177 GSI (NFS) 234290468168 Gateway to DEC-10 at York 234290524242 Radio, Space Research Establishment 234293212212 British Oxygen Company (NFS) 234299212221 Nolton (NFS) 2343 UK 234307813 Euronet Auskunft, information 234851 UK telex network 238 Denmark 2402 New Sweden 240200100110 QZ/DEC-10 Stockholm 240200100120 QZ/CD Cyber Stockholm 240200100130 QZ/Amdahl Stockholm 240200101926 KTHnet Stockholm 240200101928 QZ/Upnod Stockholm 2402001027001 DEC-10/Janus Stockholm 240200100207 CD Cyber Uppsala 240200100228 Upnod Uppsala 240200100232 BASF 7/68 Uppsala 240200100303 VAX-11/780 Umea 240200100305 VAX-11/750 Umea 240200100307 CD Cyber Umea 240200100328 NUNET Umea 240200100404 PDP-11/23 Linkoping 240200100405 TEX switch Linkoping 240200100428 TEX/Dallas Linkoping 240200100432 TEX/Dallas Linkoping 240200200203 VAX/Unix Gothenburg 240200200228 Upnod Gothenburg 240200200232 Upnod Gothenburg 240200300128 LUNET Lund 2405 Old Sweden 24050150 Scannet, Goteborg, Medicindata DEC-10 24050151 Scannet, Helsingfors 24050152 KTH/TTDS Stockholm 2405015310 Stockholm Memotec PAD QZ DEC-10 2405015320 Stockholm Memotec PAD QZ CD Cyber 2405015330 Stockholm Memotec PAD QZ Amdahl 2405015428 Uppsala Sunet Teknikum UPNOD switch 2405015432 Uppsala Sunet Teknikum UPNOD switch 2405015407 Uppsala Sunet UDAC CD Cyber 2405015404 Uppsala Sunet Teknikum Nord 100 2405015403 Uppsala Sunet Teknikum Nord 500 2405015503 Gothenburg Sunet CTH-CS VAX/Unix 2405015528 Gothenburg Sunet GD UPNOD switch 2405015532 Gothenburg Sunet GD UPNOD switch 2405015605 Ume Sunet UMDAC VAX/VMS 2405015607 Ume Sunet UMDAC Cyber 2405015607 Ume Sunet UMDAC Umdnet switch 2405015607 Ume Sunet UMDAC Umdnet switch 2405015704 Link ping Sunet SSRC PDP11/23 2405015705 Link ping Sunet SSRC TEX switch 2405015728 Link ping Sunet SSRC TEX switch 2405015728 Link ping Sunet LIDAC TEX switch 2405015732 Link ping Sunet LIDAC TEX switch 2405015828 Lund Sunet LDC Univac 24050160 Helsinki CP9500 HYLK B7800 24050200 Stockholm Scannet KI N10,Amdahl 24050201 Scannet, Kopenhamn 24050202 Stockholm Tandem Computers 2405020303 Stockholm Sunet KTH/NADA PRIME 2405020304 Stockholm Sunet KI NORD-10 2405020315 Stockholm N200 g/w DEC-10 2405020326 Stockholm Sunet US/ADB UPNOD switch 2405020328 Stockholm Sunet QZ UPNOD switch 24050204 Stockholm Prime Computers 24050205 Stockholm Televerket, testutrustning 2422 Norway 2422021001D00 Troms UNINETT UiT Cyber 171 2422021001D81 Troms UNINETT UiT NORD-10 2422021001D82 Troms UNINETT UiT NORD-100 2422021001D83 Troms UNINETT UiT NORD-500 2422021002D11 Trondheim UNINETT UNIVAC 2422021003D00 Bergen UNINETT UNIVAC 2422021005D03 Oslo UNINETT UiO DEC-10 2422021006 Oslo Scannet NSI 242211000001*D03 Oslo DEC10 24223000001*D00 Kjeller 24224500001*D00 Bergan cyber 242253000001*D11 Trondheim 242265000001*D81 U Bass Tromso 244 Finland 2624 Germany DATEX-P 26245228040187 Bonn Vax coloured Book 26245211040001 Teletex Hans Zschintzsch 26245300021711 S 7.880 BS3000 E40 HMI 26245300021712 S 7.880 BS3000 E40 26245300021713 VAX11/780 VMS (DDD pass H) 26245300021714 PDP11/40 RSX11M 26245300021715 PDP11/45 RSX11M 26245300021716 PDP11/60 RSX11M 26245300021717 LSI-X25 cableshare 26245300021719 PDP11/60 RSX11M 26245300021720 PDP11/40 RSX11M (>HEL 200,200 >BYE) 26245300021721 VAX11/780 VMS 26245300021731 PDP11/70R RSX11M 26245300021732 PDP11/70L RSX11M 26245300021733 PDP11/44 RSX11M 26245300021751 PDP11/40 RSX11M 26245300021753 PDP11/40 RSX11M 26245300021755 PDP11/60 RSX11M 26245300021756 PDP11/44 RSX11M 26245300021771 S300-R30 26245300021772 PDP11/44 RSX11M 26245300021773 PDP11/40 RSX-11M 26245300021774 S300-R30 26245300020305 CD175 NOS BERNET 26245300020106 CD835 NOS BERNET 26245300020204 CD835 NOS BERNET 26245300020113 S 7.541 BERNET 26245300020114 S 7.551 BERNET 26245300040033 LSI/X25 SOFTEC BERLIN 26245400090560 DESY VAX EMBLHH 26245724740001*D S 7.700 KARLSRUHE FIZ-TECHNIK 26245711040006 PDP11/44 Stuttgart RUS 26245761040061 ? Freiburg UNI-Freiburg 26245890040010 VAX11/750 Muenchen Uni-Muenchen 26245400040063 PDP11/44 Hamburg DEC-Hamburg 26245615125100 S 7.700 Darmstadt GMD-Darmstadt 26245890040103 Munich Euronet 26245890040220 AMDAHL 470 Garching IPP-Garching 270 Luxembourg 272 Ireland 272431540002 Eurocom Dublin 272431540003 Eurocom Dublin 27243159000630 Eurocom Dublin 3020 Canada - Datapac 30202210002000 QL Systems (Ottowa) 302031400045 PDP11/70 CABLESHARE 302055500010 Montreal Uni Montreal 30204110001400 Comet (Canada) 30205560001200 FRI Information Services (Montreal) 30206720004000 University of British Columbia (Verex on TI) Amdahl MTS 30206720004200 University of British Columbia (Amdahl 470 MTS) 302083200013 TRIUMF VAX Vancouver 3025 Canada - Globedat 3029 Canada - Infoswitch 3103 USA - ITT 3104 USA - WUI 3106 USA - Tymnet 3106*DINFORMATION 3106*DDISDCUSFSA HONEYWELL (TTUB124 CNTSS USFSA LAYER) 310600266400*D SLAC SLACBIS 1200 3106*D DISDCUSFSA TTUB124 CNTSS USAFA LAYER Honeywell 31060050 National Library of Medicine 31060051 SDC Search Service 31060052 Marc Analysis Research (Massachusetts General Hosp) 31060052 Massachusetts General Hospital 31060053 GTE Laboratory 31060054 Ohio State University 31060055 Honeywell, Phoenix Arizona 31060056 Data Resources Inc 31060058 State University of New York 31060059 National Library of Medicine 31060060 Battelle Memorial Labs 31060061*DLRRS;IPSSLON Lockheed Information Systems (JHGF3740) 31060063 Stanford Library Centre for Information Processing 31060063 Standard Centre for Information Processing 31060065 ESA (European Space Agency) 31060066 Bloodstock Research 31060067 OCLC Incorporated (Ohio College Library Centre) 31060068 Stanford Medical Centre 31060068 Sumex Computer Project at SRI 31060069 Young & Rubican 31060070 Johnson and Johnson 31060071 Cornell University 31060073 Data Resources Inc (300 bps) 31060074 Honeywell Information Services Datanetwork (300 bps) 31060075 Honeywell Information Services Datanetwork (110 bps) 31060076 Honeywell Waltham 31060077 Federal Judicial Centre 31060078 Wright Patterson Air Force Base 31060080 Standard Oil of California 31060081 Computone Systems Inc 31060082 Utility Network of America 31060083 Comnet 31060084 Optimum Systems Inc 31060086 Computer Usage Company 31060087 Proprietary Computer Systems 31060089 Information Systems Design 31060092 Phelps Dodge Corp 31060094 Chrysler 31060096 Remote Computing Corp (Ross Systems) 31060097 STC Timesharing Services (Storage Technology Corp) 31060098 Lockheed Jet Plane 31060099 Dialcom Inc 31060101 Sun Information Services 31060102 National Library of Medicine - Rutgers 31060102 Rutgers University 31060103 The Computer Company 31060104 Data Resources Inc (300 bps) 31060105 Secobi-Conacyt Mexico 31060106 Shell Oil 31060109 University Computing Company 31060110 Chrysler 31060111 Honeywell, Phoenix Arizona 31060112 Touch Ross 31060113 Bank Of America 31060115 Commercial Union Leasing 31060116 Optimum Systems Inc 31060117 West Publishing 31060119 Rapidata 31060129 Fairchild Camera and Instruments 31060134 University of Texas Health Science Centre at Dallas 31060135 Kodak 31060136 HDR Systems 31060137 MJK Associates 31060138 Computer Resources Inc 31060143 The Information Bank 31060145 Interactive Data Corp 31060146 Citibank NA (Interactive Computer Centre) 31060147 CTI (Container Transport International) 31060149 Judicial Department 31060150 Judicial Department 31060151 Polystems Inc 31060152 Xerox 31060154 Randolph Air Force Base 31060155 IP Sharp Associates 31060156 Methodist Hospital 31060157 Fritz Computer Inc 31060158 ITEL Corp (300 bps) 31060159 Compuserve 31060160 Insco Systems 31060161 Pullman Computer Service 31060163 United Computing Systems 31060164 Sales Forces Company 31060166 Computer Sharing Services 31060168 American Computer Referral 31060169 Basic Timesharing Inc 31060173 National Computer Network of Chicago 31060174 US Geological Survey 31060175 US Geological Survey 31060176 US Geological Survey 31060177 Data - Tek 31060179 LBL San Francisco (CDC) 31060179 Lawrence Berkeley Labs 31060181 McDonnel Douglas Automation 31060182 Coordinated Management Systems 31060190 Manufactures Hanover Leasing 31060190 Stalding Corp 31060191 Dalgety Inc 31060192 Compugraphic Corp. 31060195 Marc Analysis Research (Massachusetts General Hosp) 31060210 Procter and Gamble 31060226 Anistics 31060226 Interactive Market Systems (Anistics) 31060232 Scientific Timesharing 31060242 Timesharing Resources 31060252 Computer Science Corporation 31060255 Timesharing Associates 31060276 Management Decision Systems Inc 31060288 SRI San Francisco (UNIX) 31060288 Stanford Research Institute (SRI) 31060303 Scientific Timesharing 31060307 Infomedia Corporation 31060323 TRW Defence & Space Systems Group 31060432 Interactive Market Systems 31060466 Bibliographic Retrieval Services 3106900061 Lockheed Information Systems 3106900803*DK Dialine2 (type K to connect) 3110 USA - Telenet 31102010002000 Insco Systems 31102010002300 American Information Services 31102010002400 The Information Bank 31102010002500 New Jersey Institute of Technology 31102010002800 Olcott International Company 31102010003700 Informatics Inc 31102010159200 Scientific Process & Research Inc 31102020002100 Scientific Timesharing 31102020002200 Scientific Timesharing 31102020010900 CIS Information Service 31102020014100 GT-Net Telemail 31102020119500 Gallaude College Computer Centre 31102030006400 NCSS Bureau 31102030178900 Yale University Computer Centre 31102120002000 Bowne Timesharing 31102120002500 Interactive Market Systems (Anistics) 31102120002800 Burroughs Corp (NYC data centre) 311021200141 TELEMAIL 31102120014200 GT-Net Telemail 31102120139200 Memorial Dose Distribution Computation Service 31102120140600 MAV Systems (300 bps) 31102120157800 IP Sharp Associates 31102120158000 SDL International (1200 bps) 31102120158500 SDL International (300 bps) 31102120158800 DSL Systems Inc 31102120159500 SDL International (1200 bps) 31102120162000 Telestat Systems Inc 31102120162700 Telestat Systems Inc 31102130002200 Interactive Systems Corporation 31102130002700 Mellonics Information Centre 311021300029 TRW Defence & Space Systems Group 31102130003300 SDC Search Service (300 bps) 31102130004400 SDC Search Service (1200 bps) 31102130004700 University of Southern California 31102130017000 Dialine 2 (type K to get in) 31102130021908 CALTECH VAX (VALENTE ENZO) 31102130021909 CALTECH VAX 31102130140300 Marshall & Swift Publication 311021700021 University of Illinois - Urbana 31103010002000 National Library of Medicine 311030100028 The SOURCE PRIME 31103010126500 Informatics Inc 31103030002000 Computer Sharing Services 31103030002300 Broker Services Inc 31103030113100 EDI Computer Services 31103030113200 EDI Computer Services 31103030113300 Energy Enterprises 31103050116300 Florida Computer Inc 31103120002200 National Computer Network of Chicago 31103120003100 Continental Bank 31103120003200 Continental Bank 31103120004900 American Hospital Supply Corporation 31103120107300 Commodity Information Services 31103130004000 ADP Network Services 31103130006200 Merit International (MIT) 31103130139800 Merit Computer 31103140007200 Environmental DataNetwork Inc. 31103140106500 McDonnel Douglas Automation (300 bps) 31103140106600 McDonnel Douglas Automation (110 bps) 31103140106700 McDonnel Douglas Automation (1200 bps) 31103140161000 McDonnel Douglas Automation (300 bps) 31103150002000 Bibliographic Retrieval Services 3110412014600 On-Line Systems Inc 31104140002000 A.O. Smith Data Systems Divisions 31104150002000 Lockheed Information Systems (JHGF3740) 31104150002003 Dialine 2 (type K to get in) 31104150006900 SLAC (BRYANT PAUL) 31104150008000 Harper (DEMO DEMO) 31104150048000 Lockheed Information Systems 31104150123600 Hydrocomp Inc (300 bps) 31104150123700 Hydrocomp Inc (1200 bps) 31104150126800 ITEL Corp (300 bps) 31104150126900 ITEL Corp (1200 bps) 31104150159700 Stanford Library Centre for Information Processing 31104150159700 Standard Centre for Information Processing 31105150139600 State University of New York 31105160002200 Timesharing Resources 31106020160900 Timesharing Associates 31106030002000 Dartmouth College 31106030005000 Corporate Timesharing 31106030154700 Raytheon Company Scientific Computer Service 31106070002000 Cornell University (134.5 bps) 31106070002100 Cornell University (300 bps) 31106070002200 Cornell University (1200 bps) 31106070002300 Cornell University (1200 bps) 31106070003600 TIPO Computer 31106080002500 University of Wisconsin 3110608016630 University of Wisconsin 31106120002500 Honeywell Information Services Datanetwork 31106120002700 Honeywell Information Services Datanetwork 31106120106500 Honeywell Information Services Datanetwork (300 bps) 31106120106500 Honeywell Information Services Datanetwork 31106120106600 Honeywell Information Services Datanetwork (110 bps) 31106120106700 Honeywell Information Services Datanetwork (300 bps) 31106120106900 Honeywell Information Services Datanetwork (134.5 bps) 31106170002000 Bolt Beranek & Newman 31106170002300 Computer Corporation of America 31106170002400 AVCO Computer Services 31106170003600 Data Resources Inc 31106170006700 Management Decision Systems Inc 31106170007000 Interactive Science Corp 31106170007600 Interactive Science Corp 31106170008000 III Systems Inc 31106170012000 Cullinane Corp 31106170013700 Massachusetts Institute of Technology 31106170013800 Massachusetts Institute of Technology 31106170013900 Massachusetts Institute of Technology 31106170014000 Massachusetts Institute of Technology 311061700611 csnet-sh at BBN it is on ARPA 31106170101600 Data Resources Inc (300 bps) 31106170101900 Data Resources Inc (300 bps) 31106170116100 First Data Division/ADP Inc 31106170116200 First Data Division/ADP Inc 31106170125800 Data Resources Inc (134.5 bps) 31106170126900 Interactive Management Systems 31106170127500 Massachusetts Institute of Technology 31106170139000 Massachusetts Institute of Technology 31106170140300 Massachusetts Institute of Technology 31107030002000 Litton Computer Services 31107030002100 American Management Systems 311070300056 PRC Computer Centre Inc 31107030505200 Digital Broadcasting Corporation 31107130108300 Corporate Services Inc 31107130156500 Rice University 31107140002000 Science Applications Inc 31107140113700 Engineering Supervision Co 31107170002000 Brodart Inc 31108010113700 Environmentech Information Systems 31108040002000 Multiple Access Computer Group 31108080001046 UKIRT 31109140002200 Electronic Tabulating Corporation 3113 USA - RCA 3125 312541500007*DK Dialine 2 (type K to get in) 4401 Japan 44015512134*D8 505 Australia 505273720 000 University of Queensland 5053003410 CSIRONET Sydney 748 Uruguay
SERVICE NET NETWORK SERCNET * * RAL PSE1 * 1 RLIA XIF %10000001 Rutherford Laboratory IBM system 1.XXX RLIA PS %00000000 2 RLIB XIF %100100000 IBM 3081 VM/370 at Rutherford 2.XXX RLIB PS %00000000 2.XXXP RLIBP X %00000000 RAL IBM full screen 3270 emulator 3 RLIC XF %10000001 RAL IBM MVS 5 GWGA QXSMIF %00101000 GEC 4180 at Glasgow 6 RLPA XSIF %00110000 PRIME 750 at Rutherford (Prime A) 14 NMPA XSIF %00100000 PRIME 400 at Nottingham 18 ZUGA QXSMIF %00101000 UCL ICF GEC 4085 E.W.S. 20 RSGA QXSMIF %00101000 Appleton ICF GEC 4070 22 RLPB XSIF %00110000 PRIME 750 at Rutherford (Prime B) 23 RLPC XSIF %00110000 EBL PRIME 750 at Rutherford (Prime C) 25 WKPA XSIF %00100000 PRIME 550 at Warwick 26 SYPE XSIF %00100000 PRIME 550 at Surrey 27 SVPA XSIF %00100000 PRIME 550 at Sussex 28 UMPA XSIF %00100000 PRIME 750 at UMIST 32 ZUXA XSF %00110000 UCL LSI11 Gateway 33 RLVS XSF %00100000 Starlink VAX 11/780 at Rutherford 36 XXVA XSF %00100000 HEP VAX 11/780 at Oxford * 37 RLGK QXSMIF %000111000 ICF Development GEC 4090 at Rutherford 4010 RLGK F %100111000 GEC 4090 via YBTS ftp [NODOC] * 4011 RLDD XI %00000000 RADSTATE PDP-11 4012 RLDE XI %00000000 R3 PDP-11 SNS 4400 RLNA XI %00000000 R25 Nord (EISCAT Project) 38 XXWB SIF %10100000 DL GEC 4065 Graphics Workstation, Oxford (Zoology) 40.PSS PSS PXIF %00110000 40.PSS PSS77 X %10110000 PSS with FTP-77 (not supported!) 40 RLXA QXSMIF %00111000 RL GEC 4065 PSS Gateway 46 ZMGA QXSMIF %00101000 ICF GEC 4070 at Queen Mary College 48 ZIVA XSF %00100000 HEP VAX at Imperial 50 RLIT XIF %00100000 [NODOC] 53 RLWA S %00000000 Development GEC 2050 at Rutherford 65 RLVB XSF %00100000 BCRG VAX 11/780 at Rutherford 66 ZUXB I %00000000 UCL LSI-11 TELETEX development 67 RLGM QXSMIF %00101000 DCS GEC 4190 - RAL Mail Machine 68 GWIA FXI %100100000 IBM 4341 at Glasgow 69 RLVC XF %00100000 RAL VAX 11/750 St/CB in R26 70 RLDD X %00000000 RADSTATE PDP-11 1000 RLE1 QXIF %10000000 RL GEC Exchange 1 [NODOC] 1100 RLCA XF %00000000 RL PAD ATLAS 01 1101 RLCB XF %00000000 RL PAD ATLAS 02 1102 RLCC XF %00000000 RL PAD in R25 1103 RLCD XF %00000000 RL PAD in R2 1104 RLCE XF %00000000 RL PAD in R1 (tel exch) 1105 RLCF XF %00000000 RL PAD in R1 (1.75) 1106 RLCG XF %00000000 RL PAD ATLAS 03 1107 RLCH XF %00000000 RL PAD (spare) SERVICE NPRT 1108#01 RLCIP XF %00000000 SERVICE NET 1108 RLCI XF %00000000 RL CAMTEC pad 1109 RLCJ XF %00000000 RL CAMTEC pad 1110 RLGE QXSMIF %00101000 GEC 4070 Print server at RAL 1113 RLCM XF %00000000 RL CAMTEC pad 1114 RLCN XF %00000000 RL CAMTEC pad 1117 RLID XF %10000001 UTS on Atlas 10 1200 ZIIA FXI %100100000 IBM 4341 Imperial College 1201 BHIA FXI %100100000 IBM 4341 BIRMINGHAM * * RAL PSE2 * 2000 RLE2 QXIF %10000000 RL GEC Exchange 2 2100 RLPD XSIF %00100000 PRIME 400 at Rutherford (Prime D) 2101 RLPF XSIF %00100000 Prime 750 - Technology division 2102 RLPE XSIF %00100000 Prime 550-II - Lab overheads. 2103 RLGF QMXSIF %00101000 GEC 4065 workstation in R3 2104 RLPG XSIF %00100000 RAL Tech Division Prime * 2105 RLGB QXMIFS %100101000 ICF GEC 4090 - RLGB at Rutherford 2105.XXX.STATUS STATUS PX %00000000 Status Polling process at RAL 2201 XXDA XF %00100000 Oxford HEP DEC 10 2202 RLXC X %00000000 Reverse Pad at RAL 2250 UMPB XIFS %00100000 UMIST Prime 2250 2251 BAPA XIFS %00100000 BATH Prime 2250 2300 CFGA QXSMIF %00101000 ICF GEC 4090 at Cardiff 2301 BRGA QXSMIF %00101000 ICF GEC 4090 at Bristol 2302 BHGA QXSMIF %00101000 ICF GEC 4080 at Birmingham 2303 RGGA QXSMIF %00101000 GEC 4070 workstation at Reading 2304 SNGA QXSMIF %00101000 GEC 4070 workstation at Southampton 2306 CDGA QXSMIF %00111000 ICF GEC 4085 at Cranfield 2307 LTGA QXSMIF %00101000 ICF GEC 4090 at Leicester Poly 2309 PHGA QXMIFS %00101000 SERC GEC 4160 at Polaris House 2400 EXXA XF %00110000 GEC TS gateway at Exeter *2400#04 EXXA X %00000000 GEC gateway at Exeter (X29) *2400.234239200001 EXSA F %100100000 System 4 *2400.234239200002 EXPA F %100100000 Prime *2400.234339200003 EXDA F %100100000 PDP-11 Unix *2400.234239200004 EXDB F %100000000 LSI-11 with Bristol FTP-80 2500 RLGD QXSMIF %00101000 RL ISG 4090 2501 RLGU QXSMIF %10001000 RL GEC 4065 Universe LDC 2502 LHGU QXMIFS %10001000 Loughborough GEC 4065 Universe LDC 2503 RLXU X %00010000 Universe TS gateway to ring 2504 RLGV XMIFQ %10101000 Universe Exhibition LDC 2505 RLDA XF %00100000 RAL DCS PDP 11/70 2506 RLDC XF %00100000 RAL AMPTE PDP11 2507 XXDB XF %00100000 Oxford PDP-11/70 2600 GXVA XFS %00100000 RGO VAX 11/750 50200015 XXVC XFS %00100000 OXFORD COMP. CENTRE VAX 2602 GXVS XFS %00100000 RGO STARLINK VAX 11/780 2603 MAVS XFS %00100000 Starlink VAX at Manchester *2604 RLVD XFS %00100000 RAL IKBS VAX 11/750 in R30 2605 BHVS XFS %00100000 Starlink VAX at Birmingham 2700 XXCA XF %00000000 JNT pad at Oxford Crystallography 2701 XXCB XF %00000000 Oxford University Computer Centre 2702 APCA XF %00000000 Camtec PAD at Culham Laboratory 2703 PLCA XF %00000000 Plymouth Poly Pad. 2704 DUCA XF %00000000 DURHAM PAD 2705 BACA XF %00000000 Bath PAD 2706 MUCA XF %00000000 MSSL PAD 2707 XRCA XF %00000000 Oxford Record Link Pad 2800 CDXA XF %00110000 Cranfield CPSE 2850 BRXA XF %00110000 Bristol CPSE 2900 SWXA XF %00110000 Swansea CPSE 2901 RLQA XF %00000000 CB PERQ in R30 * * RAL PSE3 * 3000 RLE3 QXIF %00000010 RL GEC Exchange 3 *3101 RLGK QXIF %100101000 [NODOC] 3560 RLGG QXSMIF %00101000 GEC 4065 workstation at RAL * * RAL PSE 4 * *4010 RLGK XF %100101000 GEC 4090 (Sys Dev m/c) 4100 RLQB XF %00000000 R30 PERQ SUPPORT 4200 RLPH XSIF %00100000 RAL TECH DIV PRIME 2250 4500 RLGR QXIF %101100000 GEC Development 6300 4600 RLVA XSF %00100000 HEP VAX 750 4601 RLVE XFS %00100000 CD VAX 11/750 (VMS) 4602 RLVD XFS %00100000 IKBS Vax 11/750 at RAL 4603 RLVF XFS %00100000 R3 SNS PUNCH VAX * * Daresbury area * 1000200*D:TSO DLIA PX %00000000 1000200*D:TSO TSO PX %00000000 1000050.064700 DLIAS PS %00000000 Daresbury central computers [NODOC] 1000200*D:NETSTAT NETSTAT PX %00000000 1000200*D:ITP.1000450.046400 TELLDL PX %00000000 1000200*D:ITP.1000450.04FE00 HELPDL PX %00000000 1000200*D:ITP.1000450.46500 NETMON PX %00000000 1000200*D:ITP.1000450.44400 LOG PX %00000000 1000050.064700 DLIAI PI %00000001 Daresbury Laboratory central computers 1000050.044500 NETSTAT PI %00000000 1000050.046400 TELLDL PI %00000001 1000050.04FE00 HELPDL PI %00000001 1000050.46500 NETMON PI %00000001 1000050.44400 LOG PI %00000001 1000050 DLXA XIF %00000000 DL Gateway to Daresbury site * Daresbury terminal concentrator 1000100 DLGC SIFX %00100001 DL GEC at Daresbury (Network monitor) 1000200 DLGD SIFX %00100001 DL GEC at Daresbury (Network converter) 1000300 DLGE SIFX %00100001 DL GEC 4090 at Daresbury 1000450 DLXB X %00000000 DL Site Area RJE4 1002000 DLVA FX %00100000 DL SRS VAX 11/750 at Daresbury 1002100 DLGM FSX %00100000 DL GEC 4065 MAIL machine 1003000 DLVB FX %00100000 DL VAX B 1004010 DLCA XF %00000000 JNT pad at DL 1004010 DLCB XF %00000000 DL Pad 1 1011700 DLGA SIFX %100100001 DL CSE/1 GEC 4190 at Daresbury 1011750 NNGA IFX %00100000 DL NSFD/R1 at Daresbury 1011750 DLGB SIFX %00100001 DL NSFD/R1 GEC 4085 at Daresbury * * Manchester area * 1020050 MAXA XIF %00000000 DL Gateway to Manchester site 1020100 MAGB SIFX %00100001 DL GEC 4190 at Manchester 1020300 MAGA QXSMIF %00101000 Manchester GEC 4090 E.W.S. * * Liverpool area * 1030050 LLXA XI %00000000 DL Gateway to Liverpool site 1030100 LLGA SIFX %00100001 DL GEC 4085 at Liverpool 1030199 LLIA SXF %00100000 Liverpool HEP IBM 4331 1030200 LLGB SIFX %00100001 Liverpool Graphics GEC 4085 1050000 BGXA XIF %00100000 UCNW (Bangor) DEC-10 Gateway * 1060000 - 1060999 is the Birmingham sub area 1060000 BHXA XI %00000000 DL Gateway to Birmingham site 1060100 BHGB XIF %00100000 DL NSF GEC 4065 at Birmingham * 1070000 - 1070999 is the Sheffield sub area 1070000 SHWA XI %00000000 DL PDP-11 workstation at Sheffield 1070100 SHGA QXSMIF %00101000 ICF GEC 4085 at Sheffield 1070200 SHXA I %00000000 DL Gateway to Sheffield site 1080100 BDGB XSIF %00100001 Bradford GEC 4065 1080200 BDGA QXSMIF %00101000 GEC 4090 at Bradford 1080500 LEVA SXF %00100000 VAX at Leeds University (Mech Eng) 1080501 LEVC SXF %00100000 VAX at Leeds * 1090000 - 1090999 is the Glasgow sub area 1090000 GWXB XI %00000000 DL Gateway to Glasgow site 1100200 WKWA XI %00000000 DL PDP-11 workstation at Warwick 1101100 LAVA SXF %00100000 Lancaster University HEP VAX. 1102000 KECA XF %00000000 Keele University CAMTEC PAD 1500100 NEDA XSF %00100000 Newcastle DCS-UNIX front-end 1500200 NEVA XSF %00100000 Newcastle VAX 11/780 1500300 NEGA QXSMIF %00101000 ICF GEC 4070 at Newcastle 1501000 DUGA QXSMIF %00101000 GEC 4080 workstation at Durham 1502000 DUVS XFS %00100000 Durham Starlink VAX 1600100 SDPA XSIF %00100000 PRIME 750 at Salford * * London area * 2002100 ZKGA SIFX %10100001 GEC 4065 at Kings college, London 2003000 ZAWA XF %00000000 PDP-11/70 at Birkbeck college 2005002 ZUVS XFS %00100000 Starlink VAX at UCL 2005003 ZUPA XSIF %00100000 PRIME 550 at UCL 2005004 ZUGU QXSMIF %10001000 GEC 4065 Universe LDC at UCL *2005100 ZUXC IFS %00110000 UCL CS LSI 11 2005200 ZUXU FX %00110000 UCL Universe T/S Gateway 2006001 ZIGA QXSMIF %00101000 GEC 4065 workstation at Imperial College, London 2007001 CYPA XSIF %00100000 PRIME 550 at City University *2008001 ZWGA QXSMIF %00101000 GEC 4085 workstation at Westfield College, London 2009001 ZMVA XFS %00100000 QMC Physics VAX. 2010001 KTDA FX %00110000 PDP 11/34 gateway at Kent * * Bidston area * 3000000 BIE1 QXIF %10000010 GEC Exchange NERC Bidston [NODOC] 3001000 BIXA X %00000000 Bidston Sub-switch 1 3002000 BIXB X %00000000 Bidston Sub-switch 2 *3003000 BIHT X %00000000 Bidston Honeywell test *3006000 GEXA XF %00000000 GEC Exchange at NERC Edinburgh *3008000 *3008100 *3008200 *3008300 *3008400 *3008500 *3008600 * * Swindon area * 4000000 SIE1 QXIF %10000010 GEC Exchange NERC Swindon [NODOC] 4008100 HQGA QXSMIF %00101000 GEC 4090 at NERC Swindon * * London area * 40000040 LNXB QXSMIF %00111000 JNT London 4160 PSS Gateway 5000000 LNXA X %00000000 Gateway to the London area 5111600 ZUXC IFS %00110000 UCL CS LSI 11 5112300 ZUVA XF %00100000 HEP Vax at UCL (Physics Dept). 5180001 RHVA XF %00100000 Vax at Royal Holloway. 5201000 ZEVA XF %00100000 Vax at Queen Elizabeth College * * York area * *(wonder who notices)6000000.ITP BALHAM PI %00000000 6000000 YKXA XIFS %00110000 DEC10 Gateway at York (BALHAM) 6000001 YKDA FXS %00100000 Comp. Sci. PDP 11/44 6000002 YKDE FXS %00100000 Psychology PDP 11/40 6000003 YKDB FXS %00100000 S/W Technology Vax 11/750 6000004 YKDF FX %00100000 Development VAX 11/730 Unix *6000006 YKQA X %00000000 Dev ICL PERQ *6000007 YKQB X %00000000 Chemistry Perq 6000008 YKDC FX %00100000 Computer Service Vax 11/780 *6000009 YKDD XF %00100000 Comp. Sci. Vax 11/750 6000100 YKXD XF %00110000 CPSE Gateway Protocol Converter 6999999 YKXG XF %00110000 GEC CPSE PAD * * Edinburgh area * 7001001.XXX GRETNA PX %00000000 7001001 EDXA XSIF %00110000 DECNET Gateway at Edinburgh (GRETNA) 7001002 EDXB XSIF %00110000 RCONET Gateway at Edinburgh 7001003 EDXC XSIF %00110000 Second Gretna 7001004 EDXD XF %00110000 ERCC CPSE 7001004#02.15000024 EDVA XF %100100000 Vax in Physics Dept 7001005 EDQA X %00000000 ERCC PERQ 7001010 EDCA XF %00000000 ERCC PAD 7001100 EDGA SIFX %00100000 NSF GEC 4065, Physics, Edinburgh 7001200 EDGB QXSMIF %00101000 GEC 4060 Workstation, Physics, Edinburgh 7002001 REGA QXSMIF %00101000 ICF GEC 4090 at ROE 7002002 REVS SXF %00100000 Starlink VAX at ROE 7003001 HWGA QXSMIF %00101000 GEC 4180 at Heriot Watt University 7004001 EKVA XSF %00100000 East Kilbride Kelvin Lab VAX 7006001 DDXA XFIS %00110000 DECnet Gateway at Dundee 7007001 SAVB XFS %00100000 ST. Andrews VAX 7008001 SGVA XFS %00100000 Stirling University VAX 7009001 ABXA XF %00110000 Aberdeen CPSE 7010001 SRXA XF %00110000 Strathclyde CPSE 7011001 GWXA XF %00110000 Glasgow CPSE 7012001 PAVA XFS %00100000 Paisley VAX * * Cambridge area * 8000000 CAE1 QXIF %10000000 GEC Exchange at Cambridge [NODOC] 8001001 CAGA QXSMIF %00101000 ICF GEC 4090 at Cambridge 8002010 CAGU QXSMIF %10001000 Universe GEC 4065 at Cambridge 8002011 CAXU FX %00110000 Cambridge Univ Comp Sci. T/S gateway 8002020 CAXA XS %00000000 X29 Gateway to Cambridge Data Network 8003001 CAVA XSF %00100000 HEP VAX at Cambridge 8004001 GRGU QXSMIF %10101000 Universe GEC 4065 at Great Baddow 8004002 GRXU XFS %00110000 CPSE gateway and support system 8005001 CAVS XFS %00100000 Starlink VAX at Cambridge 8006001 EAPA XSIF %00100000 PRIME 550 at East Anglia 8006002 EAVA XSF %00100000 East Anglia (Stocker) VAX 8006003 EAVB XSF %00100000 East Anglia (CRC) VAX 8006004 EAVC XSF %00100000 East Anglia (CSA) VAX * * Cern area * 9000000 CPE1 QXIF %10000010 GEC Exchange at CERN [NODOC] 9001001 CPXA XS %00000000 Cernnet Gateway 9001002 CPGA SFXMQ %00101000 GEC 4160 workstation at CERN 9001003 CPXB SX %00000000 CERN reverse PAD (Test) 9001004 CPVB SFX %00100000 CERN Hyperon VAX 9002001 CPVA SFX %00100000 CERN Development VAX 9002002 CPVC SFX %00100000 CERN Omega VAX 9002003 CPVD SFX %00100000 CERN Merlin VAX 9002004 CPVE SFX %00100000 CERN Delphi VAX 9002005 CPVF SFX %00100000 CERN Aleph VAX 000009003001#0 CPXC X %00000000 CERN Memotec Pad. 000009003002#3 VXL3 X %00000000 * * DESY PSE 9501000 DYGA QXSMIF %00101000 GEC 4160 workstation at DESY, Hamburg 9501001 DYVB SFX %00100000 Tasso VAX 11/780 at DESY, Hamburg * * UMRCC PSE * 10000001 MRXA XF %00110000 CPSE at UMRCC 10100001 MAVG XFS %00100000 VAX 11/750 at Manchester CGU * * BELFAST PSE * 11000000 QUXA XF %00110000 CPSE at Queens, Belfast 11200250 QUVA XF %00100000 Vax in Applied Maths * * * * The format of this file is as follows * * If the first character is a * then the record is a comment * If the first 7 bytes are NETWORK then this defines the network * for all subsequent records * e.g. NETWORK SERCNET * * If it is SERVICE then this defines the MAFS service used to * access this mnemonic. This is only useful for FTP (I guess) * e.g. SERVICE ASP * * All other records define translations and they all have the same format * * DTE number comes first (with possibly a TS subcomponent) * Mnemonic next (Alphanumeric) * Selection bits (and some info) * Options * rest of the record is the machine title * * e.g. 37 RLGK QSXMIF %00101000 Dev 4090 at RL * * 37 is dte number * RLGK is mnemonic * QSXMIF are the protocols it supports, see what each letter means below * %00101000 is the option halfword. Code accepts any valid number format * i.e. could have been @0028 (but easier in binary) * Rest is (obviously) the m/c title * * * Note that great care should be taken with non-alphanumeric characters. * All non-alphanumerics except _ are treated as delimiters, however, # * and - are special delimiters. # is the subaddress delimiter, and causes * the preceding DTE address to be expanded to 12 digits and the following * string to be concatenated with it. The character - is the concatenation * delimiter and causes the preceding string (after possible translation) * to be concatenated with the following string without a delimiter. * * * * Protocol letters as follows: * * P => the translation contains a process address (not TS gateway address) * e.g. 32.UCL-CS ZULA F %0 UCL service PDP-11 * e.g. 4.333 STATUS PI %0 * Q => this mnemonic and translation may be used for QTP * X => this mnemonic and translation may be used for XXX * S => this mnemonic and translation may be used for STP * M => this mnemonic and translation may be used for MTP * I => this mnemonic and translation may be used for ITP * F => this mnemonic and translation may be used for FTP * D => this mnemonic is a DNIC (useful for PSS gateway) * * The bits of the option field are as follows: * bit 15 set => backspace is implemented locally (ITP only) * bit 14 set => QTP should use the comment field for data * bit 13 set => no fast select available (PSS gateway only) * bit 12 set => STP can issue a GET for a reply buffer * bit 11 set => this mnemonic is from a TS gateway (e.g. PSS) * bit 10 set => this host supports new style mail * bit 9 set => the meaning of this name is site dependant * bit 8 set => if this bit is set then use FTP77. * bit 7 set => this site prefers TS29 to X29 * bit 6 set => this site may not issue FTP requests to this host * bit 5 set => don't put out "call connected" message (pss gateway) * *
Dear Francisco,
You will have seen from the Euronetworkshop papers that Peter Linington will be giving a rapporteur talk on national academic networks and Birgitta Carlson a talk on international European academic networks at the coming meeting in Luxembourg. I would be very grateful if you, as the Spanish representative could supply them each with a paper on the Spanish scene. The national paper could contain your plans that we talked about on your visit here and I guess that the international paper should contain comments on EARN. I expect there are some other activities of interest going on. They are looking for very short factual papers from which they can extract items for their talks but the full texts would be made available to delegates. The object is to give delegates a good idea of all the major academic network activities across Europe. I guess half a dozen pages would be an absolute maximum. Could you send the national paper to Dr. Peter Linington who is at the same address as myself and the international one to Birgitta Carlson at:-\x2\E QX Computing Centre, Stockholm University, Stockholm, Sweden.
I hope the second day of your visit was useful and that you had a pleasant return. I had several further talks to your colleagues at Networkshop and I believe they had a very useful time talking to the many delegates there. I am afraid that they must have found quite a bit of the event confusing since most of the delegates were well educated on UK academic network activities and the speakers tend to assume a high level of knowledge. None the less I think they learned a lot particularly on how we organize ourselves.
I gather your colleagues has a strong sales pitch on their visit to CAMTEC. In fact the sales talk was continued at Networkshop. John Marshall of CAMTEC went so far as to offer you the loan of a PAD plus a visit from an engineer to set it up. The catch in the offer is that they would like you to help them to get the PAD approved for use on the Spanish public packet switched network. I think CAMTEC see the emergence of a Spanish academic network as a good opportunity to extend their sales area into Spain. As a good European and Britisher I am very keen to see good European and UK companies succeed particularly when their products are of a high standard. On the other hand as a good academic I believe that you should attempt to encourage the use of the best possible equipment on your network and would not like to encourage the use of CAMTEC equipment unless it is well fitted to your need and its use does not damage your own industry. However, I would like to encourage you to accept the, not altogether altruistic, offer of CAMTEC as in my opinion they do produce very cost effective equipment which is well suited to academic use. They are also a relatively small company and can be persuaded to modify their equipment to your needs. If you do eventually decide to use their equipment make sure that you negotiate a good financial deal with then on account of being academics, bulk deals and the help you may give them in penetrating the Spanish market!
Peter Linington and myself would value a copy of any report or proposal you make regarding a Spanish Network to ensure that we can take you into account in our international activities.
I have given Manuel a receipt for the Newworkshop payments.
I look forward to seeing you in Luxembourg and do not hesitate to contact me if you require any further help.
With best wishes,
Paul Bryant.
Dear Francoise,
I am sorry that you were unable to attend the EARN Technical Meeting at Coseners House. Nuyens tells me that you would like attend the EARN ISO Transition meeting. I expect you are aware that CEPT is insisting that EARN migrates to using ISO protocols and public networks within the next few years. The purpose of the EARN ISO Transition meeting is to try and define how this migration should take place. I am expecting people who come to have a good knowledge of IBM and ISO network methods. I suspect that there are very few of us with such knowledge but it would be useful if you could study what you can find and give the migration to ISO protocols some thoughts.
I am trying to set up the meeting for May 21/22/23 at La Gaude with travel on the 20 and 24. The aim will be to produce a proposal in a single meeting rather than stagger it over several months and several meetings. I am hoping that the Association will pay for travel but I do not think that they will pay for accommodation.
May I ask if you still wish to come and if so is this dependent on funding from the association or IBM. Also, are the dates suitable.
I am hoping that Mats Brunell (Sweden), Michael Walsh (Ireland), Olivier Martin (CERN) and Roland Wolf (Germany) will be coming. IBM will provide experts.
By the way do you have a mail address I could use to get to you.
With best wishes,
Paul Bryant.
Round about 250 souls turned up. Mostly the well known networking faces. The workshop is certainly aging.
Not a lot of excitement this time but we were perhaps thinking about migration to ISO protocols with mixed feelings of horror and 'we've been here before'. Certainly the finer points of Coloured Books were forgotten and X400 and Ethernets were more interesting.
This year there were rather a lot of presentations from suppliers. Some were sales talks we could do without and some of them could well have been replaced by more important topics. There was even a presentation and demo to do with Econet which is proprietary
Mike Wells gave a review of the year. He is still claiming that the tariff structure of BT precludes its use although I thought he said it with a little less conviction. It would be interesting to see costings based on the current traffic and with the current costs which are now more apparent. It was a somewhat muted performance for a rallying call to the troops.
Ian Smith talked about the JANET progress. At last several switches are now on GEC software and several 64K lines are in place which is a cause for rejoicing. It does nothing to improve ones view of GEC's performance since GEC have been producing this software since 1978/9 (I think) and the code still lacks some of the features of the SERC software and in addition it seems the management software looks to be a long way off which was the main reason for adopting it. There are still few statistics but it seems that 1.5M bytes are shifted each day. The performance levels of 2000 hours MTBF and 4 hours MTTR are not being met. NRS is expected to be in full production later this year. This will be on a small dedicated PRIME direct on JANET. It will be further developed. The second PSS gateway's performance was disappointing. There was a small problem when the number of accounts on the gateway went over 255 (guess the problem). Authorization is still expected to be based on calling addresses. There were some problems with clearing codes from local area networks.
Peter Linington talked about the ISO transition. He went over the new ISO committee that although important to understand is boring. The Session layer and FTAM had slipped a bit. VTAM is some way of as is JTP. JTP needed the CASE protocols which are being developed. The new CCITT standards are soon to be available as a 'Red Book' which is reputed to be 10,000 pages. The computer board have now put out a new statement of requirements for networks which allows OSI as an alternative to Coloured Books. CB is looking for statements on ISO plans from manufacturers. The Transition Group has had a number of meetings and is tending to farm out much of the work to other groups such as Ethernet, CTIG, and so on. It is still the aim, now reinforced, to do the transition on an application by application basis rather than level by level. A big problem is the Yellow Book to ISO Network service (CCITT 84). There are problems with the addressing and some loss of explanatory text. Quite a number of options are available but non is ideal and we must get it right. X29 has few problems if sites first migrate to accepting 1984 and only generate it later. CTIG have rejected SS29 and for the community there seems no need for NS29.
Ian Valentine from ICL talked about X400. There are, in fact at least 10 standards:-
X400 Service and model X401 Subset of basic and essential features X420 User agent to user agent protocol X411 Message transfer agent to message transfer agent protocol X410 Reliable transfer service layer 4 to 6 X410 Mapping to T62 TELEX X408 IA5/TELETEX/VIDEOTEX conversions X409 Presentation transfer syntax X411 Remote user agent from message transfer agent X430 Access rules from teletex terminals
Not an inspired talk as it gave very little flavour of the protocol and gave not a lot of idea of the problems of bringing it into use.
BT will not be upgrading their Plessey X25 switches to X25 84 but will be putting in a new network based on Bell equipment (the British industry fails again). He went over some of the new facilities. Again, fact rather than insight.
Alan Hunter talked about the long expected Screen Management Protocol which looks like it is reaching fruition. The system supposes an intelligent box between the PAD and terminal. It is very like the full screen 3270 Cifer and unlike the 3270 Yale code which puts the intelligence at the host end and so pays the penalty of having to go into character at a time mode over an X29 link. The protocol only allows one rectangular area for writing to and so would have difficulty in allowing 3270 look alike unless a high penalty is paid as one moves from field to field. it is going to be interesting to see how it catches on. My view is that it will have difficulty catching on unless it can be used with the current popular editors such a XEDIT. The intelligent box requires very few facilities from the terminal. These are:-
Write character Partial scroll up/down Erase display Erase right of cursor Delete character Insert spaces.
The scheme makes a very interesting contrast to the full screen Cifer.
Paul Kummer reported on the Ethernet group report. The idea to use X25 over Ethernets has been endorsed at the Ottowa ISO meeting. This meeting decided on a single stack of protocols for connection style and a single stack for connectionless. None the less there were still a lot of mutterings that class 4 would still triumph.
Alan Chambers went over the history of ATS since 1981. He has almost finished a reference implementation in PASCAL which is up to transport service. Edinburgh (Pat Morrison) compiling it for an IBM PC. The code will be freely available from the JNT. The code lacks management code, buffer code and few other bits which are machine dependent. The code should be about 8K to 12K. Blue Book is being done by Westfield and is likely to be 15K. No one is yet doing X29. CAMTEC are doing a gateway in the PAD.
Andrew Hinchley talked about the ACORN ECONET to X25 gateway. It was surprising that unauthorized protocols were allowed to be mentioned! He gave his views on how to map the ECONET packet level to X25 as well as some of the higher levels. The implementation on the BBC micro seems pretty hairy as it requires a chain of processors. There was an extensive demonstration of the system afterwards which I missed.
Andrew Cole from Leeds gave a talk on computer based learning in a network environment. An interesting point that came up was whether a good man machine interface to a remote machine from a micro should be based on XXX or FTP. If FTP implimentors had been interpreted Blue Book as a file access protocol then it might well have been a good basis rather than XXX.
Vic Toy from CAMTEC gave a tutorial on the physical interfaces which are oft quoted but little understood. He was timed out at half time just as it got interesting but carried on by public acclaim. The content is too detailed to describe here but a paper is available. Clearly the old V24 is very dated but much used. Much better standards are available but are being adopted very slowly. As an aside I get very annoyed with speakers who do not keep to time just as I get annoyed with chairmen who stifle discussion. Curiously at this networkshop there didn't seem to be a lot of good discussion.
Peter Black talked on the operational support of JANET. This was centered on the better console he had developed. Whilst it was a good piece of work it is a terrible indictment of GEC that they cannot provide a good operator interface. Why should we have to pay for a so called manufacturer product and then pay to make it usable? Peter also discussed the evaluation of some line monitors which was also a good piece of work.
Michael Agnew from LDR gave a talk on 'designing for performance'. It was a bit rough. His methods were very bad as he was merely proclaiming the delights on the high speed processing executive they had built. They made the processes mirror the level structure which may well be neat but is hells inefficient on many machines. Only machines like the GEC4000 can get away with that scheme efficiently. This was just a sales talk we could have done without.
Peter TYE from ULCC described how JTMP was getting on at ULCC. The system is so complex that I will not attempt to describe it. There is now quite a bit of JTMP going on with some measure of success. Chair is still unfortunately with us although Blue Book is catching up as the underlying file transfer. It is lucky that ULCC is using our NCP. It might well be time to have another look at JTMP but not, I guess, until the 4705 is in. There are now 6 Chair sites and 7 FTP sites working into LONDON. 400Mbytes come in by Chair and 4Mbytes by FTP (over what period?). JTMP takes 13% of the machine in prime shift but 9% overall. It takes 1 Mbyte of code and 1 3/4 Mbytes of other bits which is 2 3/4 Mbytes in all. It might be possible to get quite a bit more efficiency with some tuning particularly in the movement of data in the machine.
Urgent discussions on a secret topic made me miss the ISDN/IDA Tutorial.
John Ayers from STC considered ISDN further. ISDN comes from the voice area and is based on real fast set up digital circuits. ISDN comes from CCITT study group XI and XVIII and there is lots of confusion over ideas and terminology. The standards are Q920/921 and Q930/931. The ISO standards are DIS 8348, 8348 and 8348. The interesting point with ISDN is that the data is separated from control. If JANET is to interwork with ISDN then there will be addressing problems. The packet switched camp could well learn from ISDN and visa versa.
This year there were no range discussions (a good change) but topic discussion were held and reported back on later.
Directory services for humans were needed. Should these be based on one central directory or should it be distributed? A central service would be vast with 250,000 entries although it was thought it could just about be maintained. It could only be maintained automatically by the users. The other is to distribute with each site having a directory just as Rutherford has (at least for computing division). Again it would be useful if the users could resister themselves.
The upshot of the discussion on the PABX was that there was no use for it in digital networking as it was too expensive. There could be some economies in using the wires for both data or speech and with liberalization this was possible.
John Larmouth led or rather gave the discussion on global network addressing. Since I did not have the documents in front of me I got left behind with 90% of the rest of us. It is important to get it right if we are not to have an almighty cock up.
The National Centres discussion seemed merely to say that they were needed and should have good network facilities. Conferencing came up with not a lot of enthusiasm.
John Atkins talked about the Alvey high speed network. This was based on a very curious blend of semi ISDN and packet switching. It seemed neither fish nor fowl. There will be two switching centres, using Marconi ACE switches, which can switch 2 meg Megastream links. They were going to build a box to control the switch. The packet network is to be based on two GEC 4190 switches which would use the semi ISDN switches for getting channels on a semi permanent basis. I was not all that impressed with the ideas and would have thought basing it on a full ISDN would have been better.
RACE is Research for Advanced Communications in Europe and is set up by the EEC telecoms action program. It was described by Eric Dobson. It involves 222 man years for the definition which will take a year. It will provide Europe wide broad band facilities. The full program will take 5000 man years.
There was a CR82 discussion which aimed flack left right and centre. It is clear that the Ring is not out of the wood yet and we are doing well to avoid doing any more with it just now. There is no work going towards the production of an ISO Ring yet- mores the pity.
The next Networkshop will be in Lancaster - April 14-16 1986.
Dear Mike,
John Taylor has passed me a copy of 'Condition 2' for the license for comment. I have also spoken to Vince Hathaway who provided the initial input.
Condition 2.1- This is completely acceptable.
Condition 2.2- This condition requires that we migrate to X25 as defined in CCITT 1984 within 12 months. This is unlikely to be possible. There are two reasons. Firstly there is unlikely to be any networks operating with CCITT 84, in particular public ones, in the next 12 months and secondly it will probably take longer than 12 months before the relevant work can be completed. It seems sensible to interpret the condition as requiring me to apply for an extension to the license every 12 months to give DTI an opportunity to asses progress. I would like to assume that DTI's agreement to an extension would not be unreasonably withheld as long as EARN was making sensible progress to the use of the protocol.
Condition 2.3- This condition is acceptable. Clearly it is impossible to comment on the impact of this condition until specific requirements are defined. I guess the object is to attempt to encourage us to harmonize with other activities particularly in the academic community. It is, in any case, my intention to encourage such harmonization and as the EARN Technical Coordinator I am in a good position to ensure it.
Condition 2.4- This is merely a definition of 'Open Systems Interconnection' which agrees closely with my own.
John Taylor suggests we have an early meeting to finalize the document. Should you think this necessary I would be happy to meet you at your earliest convenience.
Yours sincerely
Paul Bryant.
Dear Francoise,
I recently sent you a letter regarding the EARN ISO transition meeting. I now understand that Mr. Nuyens wants Mr. Rene Florizone to attend the meeting. Unfortunately a message was left on my telephone answering machine over a bad line and I made the bold assumption it referred to you.
I apologies for any inconvenience this may have caused.
Your sincerely
Paul Bryant.
Dear M. Florizone,
M. Nuyens from University Libre de Bruxelles tells me that you would like attend the EARN ISO Transition meeting. I expect you are aware that CEPT is insisting that EARN migrates to using ISO protocols and public networks within the next few years. The purpose of the EARN ISO Transition meeting is to try and define how this migration should take place. I am expecting people who come to have a good knowledge of IBM and ISO network methods. I suspect that there are very few of us with such knowledge but it would be useful if you could study what you can find and give the migration to ISO protocols some thoughts.
I am trying to set up the meeting for May 21/22/23 at La Gaude with travel on the 20 and 24. The aim will be to produce a proposal in a single meeting rather than stagger it over several months and several meetings. I am hoping that the Association will pay for travel but I do not think that they will pay for accommodation.
May I ask if you still wish to come and if so is this dependent on funding from the association or IBM. Also, are the dates suitable.
I am hoping that Mats Brunell (Sweden), Michael Walsh (Ireland), Olivier Martin (CERN) and Roland Wolf (Germany) will be coming. IBM will provide experts.
By the way, do you have a mail address I could use to get to you.
With best wishes,
Paul Bryant.
Object- to name all networked 'ENTITIES' Examples of ENTITIES:- File transfer service on Rutherford IBM TELEX service Terminal use of SERC GEC at Bristol
Users need not know network addresses Sites can reconfigure networks without telling users Can produce directories of services
Guarantee unique names Allow flexibility in naming Allow user friendly names Infinitely expendable
SERC Scheme:- RLIB RLGB LLIA Problems:- Inflexible, not expendable, very restrictive
Names are in a hierarchy On a site only a local name is needed Names selected by a site Names registered centrally Names have 'contexts'
Country i.e UK Sub domain i.e AC - academic community Site i.e Rutherford Department (optional) i.e HEP Entity i.e VAX 32
Two names for each entity- long or short Long- user friendly arbitrary depth. Ex- UK.AC.BIRMINGHAM.PHYSICS.VAX 32 Short- only 12 characters after UK.AC. and two components. Ex- UK.AC.RL.IB In practice UK.AC. is not used. Site chooses site name- for Rutherford Long - RUTHERFORD short - RL Site chooses entity names.
Current names converted to new ones by inserting '.' i.e RLIB becomes RL.IB or RUTHERFORD.IB RLGB becomes RL.GB or RUTHERFORD.GB No alternative long form for 2nd part Reasoning- gives easy transition and least user upset. Other sites may do other things.
All names registered at Salford. Data base maintained. Possible to get listings. One day will be able to update tables automatically. Names have contexts. Thus RL.GB may only be registered for file transfer so that user can know he cannot use interactive.
Register all its entities. Eventually use registered names. DESIRABLE We eventually get name tables updated automatically Provide means of allowing users to look up names.
Problems of some computers- GEC Running both schemes for a time Education Maintain tables from Salford.
Present: D G House Head of system services K Benn Operations leader R Brandwood Network operations supervisor P Bryant Head of network development P Gill Telecommunications supervisor C Setford IBM Hursley R Vipond IBM systems engineer C Power IBM site engineer
The meeting was set up to discuss the installation of the IBM 3725 which is to be used in conjunction with the EARN project.
The equipment is to be shipped on 20 May. It should be delivered to D G House, Building R27, Rutherford Laboratory.
No special equipment or facilities will be required for delivery.
The 3725 is to be located in the south east corner of the communications room which has recently been cleared.
The size and clearance details were not available but the equipment was the same size as a single cabinet 3705 which is already installed. The equipment requires about 3 feet clearance on three sides. A table is required for the console which is about the same dimensions as a 3101. Final location will be determined by R Brandwood.
The console will be located near the 3725 as this is most convenient for maintenance. The console can be used for a number of diagnostic purposes. Operator training will be provided by IBM. A duplicate console or the relocation of the console may be considered at a later date.
The power requirement was unknown but small and single phase. It will be supplied from board 'T'. Power connections will be organized by K Benn.
The cable will be temporarily connected via the 'T' bar. The cable being supplied is not long enough to reach the channel. eventually a longer cable of 200 feet will be installed and the 3725 re cabled.
The 3725 will be on a separate byte channel.
The 50 ft. cables supplied are not long enough to reach the patch panels and they will be lengthened as required by P Gill. Two connections are required to the already installed modems.
Software will be ordered by R Vipond but may not arrive in time. R Vipond will ensure that software will be found from somewhere on delivery.
The software will be loaded from VM. It will be configured to drive two 9.6K bisync lines.
A Burraston will organize tests with Dublin and CERN to check the final system.
C Setford offered to provide software ans setting up assistance from Hursley should this be required.
Rutherford will meet the maintenance costs and this will be organized by D G House. This will be on single shift basis.
A draft copy of the EARN agreement to be signed with IBM was considered and informally agreed. IBM will now supply a final copy for signature. A previous draft has already been agreed with G Manning.
It is now believed that a signed copy of the DTI license is on its was to Rutherford. When this has been received BT will be requested to reconnect the CERN line and give permission to use both the Dubin and CERN lines. This is likely to take a small number of weeks.
As soon as the lines are connected P Bryant will enter into negotiation with BT to decide on a tariff- possibly with IBM assistance.
At a subsequent informal meeting the problems off setting up the RSCS tables was discussed with C Setford. The problems are many but not difficult. The principle problem is to decide on a setup which will provide the right visibility of various components to the various users. In addition it must be compatible with the long term communications aims.
The mail gateway was discussed. This is now fairly well understood. The MAILER libraries must now be obtained from Columbia before any more progress can be made. It is unlikely that the gateway will be ready before mid year.
Present: B Davies (Chairman) S Rows Laser Division P Bryant (Secretary) J Hutton HEP M W Johnson Neutron Division A D Bryden Technology Division P McPherson S & A Division C Balderson Information Technology J Harrison Geophysics & Rad P Linington JNT Apologies: R E Thomas Information Technology M Hapgood Geophysics & Rad B Jones Scientific Admin
Agreed without change.
M1.1.1 New terms of reference had been produced. M1.2.1 Members had produced some input for the survey of current networking facilities. M1.2.2 P Bryant has produced a survey document of current network facilities and future requirements as RCCC/P5/85. M1.2.3 Members had produced some details of their future network requirements. M1.2.4 P Bryant had produced a summary of network requirements and these were also in RCCC/P5/85.
It was unclear whether out stations such as Chilbolton would be covered by this meeting and also what powers the meeting would have. It was considered that these points would be answered in the fullness of time.
The terms of reference were agreed.
The response to the request for information had been patchy and indicated that there was no need for a high speed local area network but only a need for a modest improvement to the current X25 facilities. In a wide ranging discussion this impression was shown to be untrue and also several other topics were introduced. There were a variety of comments.
There were likely to be a number of 50K bps links into the site soon and it was unclear what their characteristics would be and how they would be dealt with. This would be examined.
Action: J Hutton, A Bryden
The support of PADs was thought to be unsatisfactory. This was bound up with the support of printers on PADs, name registration and the high cost of maintenance of GEC computers. This would be placed on the agenda for the next meeting and members are asked to prepare comments.
Action: P Bryant
There was thought to be a requirement for 2M bps links into UTS from UNIX work stations. This needs clarification.
Action: C Balderson
M W Johnson tabled RCCC/P6/85 which described the PUNCH local area network for the SNS project. This relied on Cambridge Ring technology.
A new summary paper would be produced in due course.
Action: P Bryant
There was agreement that a site wide electronic mail service was required which was also interconnected with other mail services via JANET. This should allow mail to be sent to a 'person' by name without knowing anything about the machine he used or his login id. Computing Division has had a trial service for some time which had shown that technologically such a service was possible. The principle problems were:-
How to update the directories, preferably automatically. Where the hardware was to come from.
It was agreed that the directories should only contain members of staff and some 'friends' of Rutherford. It was not the laboratories job to attempt a UK wide service. It was suggest that the directories could be updated by the DAOs, or from the telephone directory, or from the personnel data base. The hardware should be a separate machine and not one of the current computers which is used for other unrelated tasks. Some form of directory service would be useful for use both inside and outside the site.
The meeting felt that the service should be part of the RAL infrastructure.
It was agreed to prepare a proposal.
Action: P Bryant, J Hutton
P Bryant brought to the attention of the meeting the X400 mail proposal from CERN. The proposal in the 'COMICS' report advocates a rapid move towards the X400 standard. Eventually the UK would use X400 but this would not be for some years. In the mean time gateways would be required between Grey Book and X400. This activity should not prevent the exercise to set up a site mail system.
J Hutton outlined the basic details of the Name Registration Scheme.
There are two basic problems. First the names themselves and second the implementation on the Rutherford site.
The 'names' for 'entities' have been selected and Computing Division's Network Development Meeting has put considerable effort into this. J Hutton is attempting some cooperation in the selection of names within the UK HEP community with a view to making life easier for them. He suggested that activities other than HEP may also like to encouraging some consistency in their areas. P Linington pointed out that sites were responsible for selecting names of their entities and any attempts at coordination between sites could only be advisory.
The problem of introducing Name Registration on the Rutherford site was now more difficult as it had to be coordinated between Divisions. In addition there were difficulties with the GEC computers which needed some development work. It was important to define a date for introduction of NRS. P Blanshard of Computing Division is in charge of the introduction of NRS and Divisions are asked to give him support in his work.
Action: All
It was generally agreed that a TELEX service in and out of Grey Book mail was desirable. P Bryant reported that he was purchasing a Braid TELEX system in conjunction with D Rawlinson with a view to interfacing it to a suitable computer in the network. Network Development Meeting considered that the best machine to do the work on was the GEC since it already had TELEX code which had been developed for a TELEX service over PSS. This project had been abandoned due to the outrageous tariff and the lack of an international service.
P Linington reported that Glasgow were undertaking a similar task but the front end device was a Rair Black Box.
C Balderson thought that a TELEX system from Systel was being investigated by K Jeffery's group but there were no details of this development.
It was agreed that any TELEX service must be universally available to the site and not to a restricted set. This implied that it must interface with the Grey Book mail service. Progress would be reported.
Action: P Bryant
16 July at 2pm.
M2.4.1 Examine requirements for 50 Kbps lines J Hutton into the site. A Bryden M2.4.2 Place PADs on RCCC agenda P Bryant M2.4.3 Investigate need for 2Mpbs links into UTS C Balderson M2.4.4 Prepare new summary of state of networks on the site and the requirements P Bryant M2.6.1 Prepare proposal for a site wide mail P Bryant service. J Hutton M2.8.1 Cooperate in the introduction of the NRS All M2.9.1 Report any progress on the provision of a P Bryant TELEX service.
The Division, following its several reorganizations and changes of direction, seems to have lost its way. Whilst it used to be a large expanding Division with a large number of diverse activities, it is now a contracting one with fewer tasks. An added complication is that the political and financial rules have changed. Under the new situation the Division has difficulty operating effectively and it seems appropriate to reconsider why and how it operates with a view to improving its effectiveness.
There is now considerable confusion over the mission of the Division. There seems to be two principles:-
Both of these activities are within fairly strict financial, political and complement restrictions.
In the past, the first principle was followed. Recent changes have added an element of the second whereby certain services are to the benefit of SERC and provided 'free', termed infrastructure, whilst others are charged for, such as computing services. This is confusing since the dividing line between the two is indistinct. In addition, in line with the traditions of the academic community, considerable help was given and received from other places and payment was not generally given or received. Thus the new mode of working is often at variance with the mode of working in the rest if the academic community.
The requirement to 'sell' a service has been difficult. Under such a system the customer has the choice as to how and where he obtains his resources. Clearly they may elect to purchase their own equipment or to obtain the 'free' Computer Board resources from university or National Centres. The Division, on the other hand, is constrained on the service to be provided. They can only offer services to a restricted set of customers and can only provide the service with limited complement and quality of manpower. In other words, there is an inability to bid for customers competitively and to then amend the finance and manpower to match the tasks and the income.
It seems likely that the current policy is unworkable. It is likely to result in more computers in other places. It is likely to result in loss of business for the charged machines and almost certain failure. The overall cost to SERC is likely to be high. This is because well coordinated central services for the provision of large scale computing and the coordination of effort on small machines is likely to be cost effective rather than there being small pockets of expertise scattered and isolated round Rutherford, SERC and elsewhere.
There can be several results of the anticipated failure. First, there could be a 'direction' to use the central machines. Second, there could be a return to the policy of 'free' computing, or notional charging. Third, the complete function of the division could be distributed between the divisions with Computing Division probably ending up as the computing arm of administration or High Energy Physics Division. The cynic could argue that the loss of a central computing facility could be of benefit to other divisions as it would be one less organization to deal with and would increase the money available for them to decide how to spend.
The policy of distributing computing in an ad hoc fashion is expensive. We already see that the fairly ad hoc way the VAX computers are dealt with has resulted in the total cost of system and support staff on VAXes being quite high whist the expertise on any one site is relatively low. The contrast is with GEC computers where (at lease a couple of years ago) the expertise at the centre was very high and at the sites non existent and this led to fairly low overall costs. It also led to sites being relatively unaffected by changes, or even complete loss, of staff on that site knowing that back up expertise was available. From an accounting point of view the central support looks like a high visible charge ripe for pruning- in fact such pruning was a regular event!
After Bill Walkinshaw retired the division was organized on a 'function' basis. Thus all support was done by one group and all operations by another. Before then there was a very complex arrangement whereby support of the IBM machine was on a functional basis whereas each range of small machines had a group which looked after most aspects of the range but did get some services, such as site wiring, from other parts.
On reflection one can make a number of interesting observations on the 'post Bill' organization.
For running a single very large machine some breaking of the job into functions is needed to make the task of controlling the large number of people and equipment manageable.
For small machines the policy was not a success. The reason was that there was only a small number of people working on a small machine range. Thus one found that, whereas they were run by an well integrated small team, the same amount of manpower distributed round groups was not very effective as service on the small machines tended to become the poor relation and there was a loss of flexibility in undertaking tasks. It did not seem successful to distribute the functions of one person over a set of people, often in different groups, and it became frequent that a lot of tasks were being duplicated or not done at all.
At that time 'matrix management' was in vogue. Thus it was expected that if a task was required to be done a project officer would be appointed who would have 'rights' on the resources in several groups but people would remain in their groups. The was not a success. In practice the groups were unwilling to provide the resources. One could speculate that providing such resource did not generally help the group to meet its targets, if fact the reverse was true as rarely, if ever, was any compliment added to a group to meet the added commitment. A second speculation is that the 'glory' of the successful project rained on the group the leader came from which was of little benefit to those donating staff. If fact it was a travesty of what matrix management is all about.
A better interpretation is that certain functions, such as the systems work and user support of a small machine range is best kept in a single 'vertical' group. Other functions, such as graphics, are best dealt with in a single 'horizontal' group which spans the machines. Thus related expertise is retained in a single group where it can be best exploited.
The observation is that putting all activities in a rigid organization has tended to be inappropriate for some tasks and has lowered the standard of service to the users in many cases. It has not reduced staff requirements. A tidy organization does not guarantee an efficient one.
The aim of any change must be:-
In a nutshell, a management must strive to further the happiness of its staff and customers and it is impossible to differentiate between the two.
The strategy below is based on the premise that the division is principally to provide the best possible service to the community. It assumes that the requirement to make some sort of profit is likely to fail.
A basic problem of the current structure is that functions have been placed in groups according to 'policy' rather than by the pragmatic of what will provide the best results. In a very large, expanding and well found organization this may be sensible but this is no longer the case in this division (perhaps it was never the case).
From the staff point of view, one must ensure that they feel that they have interesting, responsible and expanding jobs. This means that they should each have a satisfying blend of service and development that fits their abilities. A corollary of this is that dead end and low quality jobs should, if possible, be done away with or automated. The staff released should be upgraded in quality by natural wastage, redeployment, education and recruitment. This is clearly a long term and ongoing activity. It is also one which receives a low and 'stop go' attention. One needs to point out that staff well being appears only to be a brief and exhausting activity around promotion time, should not we groom staff for advancement continually? How often do we see public praise for any of our staff for achievements? How often to we appraise the jobs our staff are doing to ensure they are still being useful?
Recently there has been a drop in the quality and numbers of highly skilled staff. Whilst one can survive for some time in this condition it will inevitable reduce the quality of service and the opportunities and options that can be taken. For example, the loss of GEC staff means that the PSS gateway cannot be developed as one would like and it may be that the work will have to be contracted out at high cost and with no increase in the expertise in the Division, it will take a lot longer.
It is important to ensure that groups consist of people working on related topics. If this is not done there will be the 'poor relation' effect that some small areas will be swamped and neglected. This leads to a loss of morale and a poor service.
The following suggestions are intended to be illustrative of possible developments and not exhaustive or necessarily what is best.
For a long time Operations has been the technical poor relation begging for services from elsewhere. This has led to the operation of machines being primitive. The result has been the need for fairly large operating crews. A possible development would be to put effort into automating the running of the large machines to the point where they could be run unmanned on the night shift and at week ends. This is clearly not a trivial task and one which involves a lot of high grade labour and effective development of such devices as the cartridge store. It should help the reduction of less skilled staff and the increase of skilled staff.
The manpower running change control probably absorbs a man year. It used to cover all aspects of computing. It now seems only to cover the IBMs and the PRIMEs GECs and the network tend not to participate. Since it now covers a relatively small group of people who cooperate closely one asks whether it is worth the effort. One has to balance one man year of effort against the cost of the 'cock ups' that may occur.
Document production and distribution seems to have become an industry. From the writing point of view there seems to be a feeling that documentation is needed to suit the needs of the particular SERC users. In addition there is a temptation to tinker with bits of software to add some esoteric facility which then means that it has to be documented. Whist there are clearly areas where this can not be avoided there are many areas where it can. Thus one should perhaps rely more on manufacturer produced or non division produced documents. Also, if products are to be developed they should be to an already documented standard. Mail interfaces are an example area. On the distribution side, CERN appear to be encouraging users to print off 'on line' copies of documents. This is possible with high speed laser printers. Thus users can buy documents from manufacturers or print their own thus eliminating document distribution. It might be that the division should run a profit making book shop for documents which could be self financing. It is clear from a casual glance at peoples book shelves that the collection of manuals is also an industry which suggests that encouraging users only to print what they want may be cheaper.
If the division is to make a profit then the restrictions under which it works are not appropriate.
For example, since the division now has considerable expertise in PROFS then it should be possible for them to sell this to other organizations. This is impossible with the current fixed and inflexible compliment levels. Any other money making organization would tend to match its staff levels to the amount of work coming in and the finance. Also, an organization has little indication of when work will materialize and fixing staff levels on April 1 places impossible restrictions on business.
A second example is that the organization must be able to compete for business which means negotiation on prices and offering options. A take it or leave it service means that any enterprising alternative supplier knows exactly how to under cut the Division and knows the competition can not respond.
A third example is that in competing for business effort has to be put into promotion. Possible customers need to be attracted with free time to try the facilities.
The Division is about the size of a large university computing centre. On the other hand it is run on bureaucratic lines with considerable demarcation and duplication of effort.
The alternative view is that the Division should be split into small groups doing related jobs- such as graphics- and that these should be given a high degree of responsibility. The aim should be to reduce cross group meetings and other time wasting and manpower wasting activities. It should also be aimed at reducing the effort going into administration and increase the effort going into productive effort. This will need a much higher level of trust to assure people that things that concern them are being brought to their attention without the need to sit through long meetings.
In running a group it is important to ensure that the manager is in charge of the resources he has to use. It is another time wasting activity to have to beg resources from elsewhere and also to have the success or failure of ones activities dependent on the whims of someone else who may not share ones aims.
We are clearly embarrassed by our reduced circumstances. It is wrong to 'give up' and blame the system, this will not impress our lords and masters any more than it will the customers. We must be seen to be attempting to do the best possible job with what we have. The best possible job implies that we have to re-think what we are doing. This rethink should be quite independent of the people we have and the groups as they now exist. If a sensible structure can be defined then the people can be fitted into the teams and the teams assembled into the groups. The other approach of starting with staff and trying to give staff with certain grades certain levels of responsibility is doomed to failure as it ignores at its inception the object of the exercise which is - not to put to fine a point on it - survival.
Dear Sir,
Your report on the SEAS conference gives a very confused impression of the standards issues and of the European Academic Research Network.
It seems a bit late in the day for ether IBM computer users or IBM to want to influence the OSI standards. It has been well publicized for many years that the various standards bodies have been putting together a very full set of non proprietary network standards which they believe is to the benefit of the human race. You have only to read the back copies of Computing to see the many column inches devoted to the topic particularly over the last four or five years when the standards have been under heavy development. Other more specialist publications have published even more material. The standards themselves have been open to inspection at all stages of their progress by the general public. In fact, the various working parties have been quite open and have welcomed anybody who is sufficiently interested especially if they are prepared to put in effort. The various pressure groups, such as ECMA, IEEE and many others have been very active and supportive of the effort. In fact IBM themselves have contributed in many cases. Surely it is now far too late to expect to start changing standards that are, or will shortly be, ratified by ISO. Surely now is the time to be deciding on how the standards should be interpreted and exploited. To want to effect the established standards now can only be destructive and delay the date by which implementations can be made available to customers. However there is no reason why they should not attempt to influence the uncompleted standards and I am sure that their expertise will be welcome.
I certainly agree with the view that IBM should comply with the ISO standards and recent announcements by the company encourage me to believe that they are doing so and pressure from groups such as SEAS may well help the work along faster. However, it must be remembered that the standards are not yet complete enough to meet the needs of all users and it is true that currently other products, such as SNA, can provide a better service. Hopefully this position will change.
It is perfectly correct that ISO is not open to SEAS and similar organizations. It is well known that the route to ISO is via the national standards organizations plus a few other standards bodied such as IEEE and CCITT. Clearly, for better or worse, there are rules as with any activities and if SEAS wants to influence standards then they must go through the right channels. These are to start attending the various working parties and to influence then in competition or harmony with everyone else. There is absolutely no reason why SEAS or any other pressure group should not attempt to influence via the right channels but there is every reason why they should not consider themselves in a privileged position.
I find it irksome that the ISO work is only now being taken seriously when it is showing signs of success. No doubt you will recall the many people who used to claim that it could never happen and it was all a waste of time.
I must agree that there us some frustration with the difficulties of connecting to the European Academic Network due to its adoption of IBM network methods which happen to be RSCS. However, it must be added that the network is committed to migrate to ISO protocols over the next few years under an agreement with CEPT. Moreover, IBM is actively helping in this migration. As the EARN Technical Coordinator I am involved in this work and I can assure you it is being taken very seriously by both the EARN management and IBM. It must be added that it would have been impossible to set up such a network as EARN using ISO protocols and the network is providing very important services to the academic world NOW rather than in some years time when ISO products are available.
It is untrue that DTI has been isolated in not supporting EARN. In fact, in keeping with most other similar bodies they have recognized that EARN is the only immediately viable possibility and allowed it to exist whilst encouraging it to migrate to ISO protocols within the next few years. It is the academics themselves who have decided that EARN should not be established widely in the UK on the grounds that it would undermine the existing widely available academic network JANET which is based on X25 and is expected to migrate to ISO protocols in the next few years. It was decided to build a gateway between EARN and JANET at the Rutherford Laboratory which would give the 1000 or so computers on JANET access to and from the 600 computers on EARN and its associated network BITNET which is in the USA. As both EARN and JANET migrate to ISO protocols the gateway will improve in quality and may at some date become redundant.
True it has taken 20 months to obtain a license to operate EARN in the UK. This has been in the main due to the privatization of BT which came just as license negotiations were almost complete and caused them to be restarted with DTI. The flood of license requests and the low level of staff in DTI further delayed the matter. It should be added that at no time were and serious political problems encountered and certainly DTI have been most helpful in drafting a license which meets the requirements of the EARN JANET gateway.
I am surprised at the claim of Philip Dyce that the EEC was using non experts in its network activities. I have been involved with them on a number of occasions and have been impressed with the readiness they have for calling in experts from the community both commercial and academic. The ESPRIT project is heavily committed to ISO standards and the participants are well staffed with people who are very skilled and knowledgeable on the subject. They are very active in harmonizing the use of standards throughout Europe. This work is aimed at the interconnection of equipment from a variety of manufacturers and it is difficult to see how this is not in the interests of the users. If Philip Dyce is referring to the people working in the standards committees then I agree that many of them are working part time by the generosity of their employers. However, I have the greatest respect for their work and the fact that they share their time between the standards work and their work in providing network products does, I believe, lead to better standards.
The organizations which can join CEPT are well known and if IBM qualify then well and good. If not then I see no reason why they should be treated as a special case. The correct procedure, if this is important, is to get the rules changed and no doubt many other organizations such as ICL et al will join as well.
Personally I believe that in the future there will be little room for competition in communications standards. Whether there is room for SNA and ISO can be debated but if the answer is 'yes' then good gateways must be built between the two technologies. If 'no' then we must decide between the two or something else. Certainly if SNA is to be the standard then it must be taken over by ISO so that it cannot be changed at the whim of IBM or any other body not accountable to the public.
In your editorial you doubt IBMs commitment to ISO. You may be right to do so. I believe that IBM will do what it is in its interests to do. If ISO 'rings the cash register' then IBM will support it. Although IBM has a lot of power it must ultimately supply what the customers demand. Certainly the demand by the UK academics for computers to support the so called 'coloured book' protocols if they are to be purchased has led IBM to support them. If this can be achieved in a relatively small community I would have thought it even easier in the larger community represented by the EEC and other bodies.
Yours faithfully,
Paul Bryant - EARN technical coordinator.