Paul Bryant's Networking Correspondence
Computer networks is probably one of the most confusing topics of the moment. Even the experts are confused. To start to understand what it is all about an knowledge of the history is vital.
In the dark ages of the early 1970s networking was simple and easy and there was not much of it. Moreover it did not achieve an awful lot. In general you went to the supplier of your computer and he connected up a few teletypes (remember those ghastly clackity clack devices!) and if you had plenty of money he would also connect up a few remote card reader/line printer devices.
Two developments ruined this cosy situation. First- sites started to put in a variety of computers and was painfully obvious that equipment bought to connect to one manufacturers machine would not connect to another. And so one ended up with a number of expensive networks around a single installation. Second- manufacturers started to produce more sophisticated networks which would allow groups of their computers to be interconnected with a richer range of facilities. Such networks had even less chance of connecting to a variety of computers.
Whilst such a situation may well make the manufacturers happy in locking the customer into his products it did not please the customer who was faced with a high cost and a restriction on his freedom to choose the best equipment for a job.
Clearly the solution to the problem is for all suppliers to adopt a common set of networking methods which would provide a rich set of facilities. Human nature, or is it corporate nature, is not like that. Whilst everyone would pay lip service to such ideals in practice the suppliers response is 'fine as long as these methods are based on by ideas'. In other words - expecting DEC to abandon DECNET in favour of the IBM SNA network methods is just not acceptable to IBM and visa versa.
Clearly the only route forward is to attempt to define a set of non-proprietary standards which all manufacturers could subscribe to. This is easier said than done. Once you get all the parties into the same room and set them going the result is a set of standards which is bung full of options of one sort or another to meet the needs of all the participants. This does not help the cause. But more of that later. To cut a long story short such an activity was set up under the International Standards Organization and the so called Open Systems Interconnection standards have to a large extent been produced. This is quite a remarkable achievement.
So why isn't the world beating a path to the ISO standards and why is not everything sweetness and light.
The first fly in the ointment is that manufacturers have already developed their DECNETs and their SNAs and these work very well and are producing revenue. So why should the manufacturer go and develop these new animals at great expense and upset many of their users who are happily locked in. Also why allow other manufacturers to interconnect their equipment easily. Also why give up your ability to alter your protocols to meet requirements, or rather more cynically, snooker the opposition. The second fly is that manufacturers have tailored there hardware and software with great care and there existing methods do not map with great accuracy onto the ISO protocols. Foe example, it is unclear how one would operate IBM 3270 terminals across an ISO network.
So what happens, or happened, next. We, as the customers, have a choice. We make the best of what is offered or we take an aggressive approach and say we will only buy equipment which conforms to the ISO standards even though in the short term, which may be a lot longer than we would like, and we will suffer the loss in richness of the facilities in the interests of achieving interworking and forcing manufactures to support standards enthusiastically. Ideally we would want to get to the situation where it would be as inconceivable as wiring your house with anything but 230 volt equipment to buy a computer not offering ISO network facilities.
As far as the European academic community and the UK in particular is concerned the ISO pundits have it. In fact the UK was in this business a long time ago when the Coloured Book protocols were pressed on a set of reluctant suppliers. Even with reluctant suppliers the results were encouraging in that we have achieved the objectives of a network with a single set of standards which in the mail works well. There are few people in our community who would now wish to go back to HASP, Swan Net, 7020s, Metro Net and any other of the miscellany of networks which used to supply the academic needs. So if so much can be achieved in such a small community think what can be achieved on a European scale.
Well, Coloured Books throughout Europe may sound nice but these are old protocols and how much better if one can galvanize Europe into doing what we have done in the UK on a Europe wide basis but with ISO protocols. That, in fact is what is happening. An organization has been set up with this objective and it has the full support of all countries. It also has the support of the European Commission and many other governmental organizations. In fact things look promising.
To return to an earlier point there is a problem. The problem is that the ISO protocols are all things to all men and to make then really work you need to define so called Functional Standards. Functional Standards are based on the ISO ones but define exactly how a standard is to be applied. For example, if you want to transfer files then these are the protocols you will use and these are the options or the ways in which you will apply the protocols to achieve interworking. The definition of Functional Standards is now in full flight and the work has been supported by the European standards organizations CEN/CENELEC and the PTT organization CEPT. The academics are taking a keen interest and active part in this work and already Functional Standards for several areas are in draft form. It turns out that this work is far easier than the initial production of the basic standards.
So- things are looking up. Another encouraging sign is that both IBM and DEC are dedicated to providing ISO networking. In fact in the case of DEC it is expected that DECNET will eventually migrate to conform to the ISO protocols. The smaller manufacturers such as ICL and GEC have always been keen supporters of ISO and they have made similar if not stronger comments to IBM and DEC.
You may now well ask when it will all happen. Not overnight- particularly in JANET where a very large network has to migrate from its current set of Coloured Book protocols to the ISO ones without upsetting the users. Upsetting there will undoubtedly be but this has to be minimized. The migration is expected to be on a 'functional' basis. That is an application such as mail will migrate on a particular machine possibly leaving the other applications unchanged. This requires that gateways will be needed at strategic places to convert traffic to and from Coloured Book and ISO methods. This will not be easy. Estimates range from 5 to 20 years for the complete migration. Fortunately the underlying X.25 JANET network is more or less ISO and so there is no need to set up a completely new network. Already there are examples of the ISO mail protocols operating experimentally within Europe and even within JANET and the pace and extent of this work is increasing as the European's get their act together.
ECMRWF have approached me as they would like to use EARN to communicate with certain EARN users who are involved with meteorological research.
Before proceeding further I would like to be advised on if and how the requirement should be met. Unfortunately the legal status of ECMRWF is curious and the technical requirements are not absolutely straight forward.
ECMRWF is financed by 17 European countries and there appears to be no commercial involvement. The products of their activities are research and forecasts. The forecasts are provided free to the national meteorological offices to do with as they wish. It is the research activities which need to use EARN and not the forecast ones. The legal aspects of ECMRWF are strange in that they do not operate under UK law and seem to have a certain amount of power to do what they want. For example it appears they could set of a satellite dish on their site without permission from DTI!
Although ECMRWF have an IBM 4341 model 12 it would be preferable for EARN activities to be to their VAX computer. For a direct RSCS connection this would require the J-NET software to be purchased from Joiner Associates.
It is unclear whether there any objections from JNT or RAL to a direct RSCS connection between ECMRWF and the RAL IBM computers. The equipment on the IBM would be that donated by IBM and ECMRWF would meet the costs of the lines and modems. It is unclear whether RAL would wish to make other charges although such charges would not be in the spirit of EARN.
It would be possible for ECMRWF to be connected to JANET and to use Coloured Book protocols. What is the JANET reaction to this.
It would be possible for ECMRWF to connect to PSS and to use Coloured Book protocols and it is my impression that JNT and RAL would not object. This would be less satisfactory due to the problems of the PSS-JANET gateway with respect to charging and also the quality of the service may be lower.
It is unlikely that EARN will object to the use of their network by ECMRWF but their advice is sought to confirm this.
ECMRWF are reluctant to use Coloured Book protocols on the grounds that other member countries may request that ECMRWF also support their favoured networking methods. EARN is an international network and therefore will not attract the same comments (ECMRWF comment).
The license for EARN from the DTI allows connections from overseas EARN sites to Rutherford and allows traffic which originates or terminates in JANET to use those connections. Thus for a direct connection to RAL the line would have to be designated as part of JANET and therefore the exact status of ECMRWF would have to be determined and possibly a dispensation obtained from DTI. Alternatively dispensation can be sought from DTI for a direct connection from ECMRWF to Rutherford which would not involve JANET. Whether any such line used RSCS or Coloured Book protocols is irrelevant.
ECMRWF are intending to use ISO protocols when appropriate.
I seek advice before proceeding further.
Networking information gleaned from the press is confusing to say the very least. On the one hand we see adverts that this product or that will solve all your network problems. We see confident proclamations from IBM, DEC and a few more that SNA or DEC provide the network route to the future. On the other hand we read of customers asking why this product or that does not work or is not available. We see a certain amount of frustration with products which just do not communicate. We see a raging battle between SNA, DECNET and ISO with each claiming the imminent demise of the others.
So what is happening.
Reading the documents leads one to the view that there is a disease called 'network diarrhoea'. There are more buzz words per column inch than you get on an income tax form and moreover one gets the impression that the long works are aimed at fooling you (if in manufacturer blurb) or letting you know how cleaver the author is if otherwise. One may be forgiven in thinking that there is a conspiracy of confusion.
It is very clear that the majority of the buying and consuming public are confused and bemused by the whole issue.
I guess we all remember the balmy days when things were easy (perhaps we did not think so at the time) and you just got in your friendly supplier to put in his proprietary terminals and work stations and away you went.
This went well until the late 1970s when it started to dawn on customers that their favorite piece of equipment could only be used to communicate with a specific brand of computer. So for each machine they owned they had to put in a whole cobweb of lines at some cost.
It was at this time that some suppliers perceived the discomfort and started to come out with all the 'we have been in the business for 100 years and will solve all your problems' spiel. In practice their offerings, although useful were very limited and fell into a number of well known groups. These were:-
History goes on to record that the idea that a network should also switch data traffic just like a telephone network caught on, at least in the imagination of the public. It had become clear that using the actual telephone network was a non starter. The data rates were to slow and the costs were to high and you could wear your finger out dialling calls (even if it is an electronic finger).
The idea of dedicated data networks which could switch became popular. Or to put it another way everyone leapt aboard with a whole slew of incompatible products. In fact there were 3 basic lines of development:-
So- to summarize we have a plethora of network technologies each with considerable draw backs and it would be a brave man who stated which was worse because for certain as we sit here they are all pathetic.
We all want without exception ('all' refers to the poor customer) the unified networking that will suit everyone. Well- you will not get it- at least not in the foreseeable future.
To start with products such as SNA and DECNET are well tuned to use IBM or DEC products and unsuited to other uses. To take my favorite example- how would a IBM 3270 terminal fit with applications on a DEC computer and how would VT100 fit with IBM applications? The answer is badly if not very badly. So the upshot is that SNA will not rule the world any more than DECNET will. No doubt such technology could be adapted but I cannot see the will from either camp to do it. Both these technologies also have a major draw back that they are designed for a set of large computers linked together rather than a set of computers linked to a network which exists quite independent of the computers.
Both SNA and DECNET also have a drawback which is even more serious in that they are the product of a particular manufacturer and he can, at his own whim, change the protocols and snooker the opposition. And that is just not on for 'every mans' network.
So what we are looking for is a network technology which is non proprietary, allows the connection of everything to anything, and meets every ones needs. The quick answer is - impossible.
The long answer is that it is possible to some extent.
The light on the horizon is that the international standards organization ISO (incidentally one of the better results of the United Nations) has recognized the problem and is setting about producing standards. The work started slowly but is now gaining speed. At first it was thought to be a waste of time but now there is a sporting chance that it will provide a path to the promised network.
The problem that faces standards is that they must meet many requirements. They thus become very complex. Everyone wants to ensure that they will meet his particular need. Thus they are painfully slow to produce.
The remarkable fact is that many of the protocols needed are now very well advanced and implementations are being produced. In the next few sections some of the major items are considered.
To start talking about networks without some sort of model for the topic is a waste of time. So the first job for ISO was to define what they were talking about. They decided in their wisdom that network protocols had layers. The bottom layer, for example, talks about plugs and sockets while the top layer talks about applications such a file transfer. In between all sorts of other thongs happen. In fact there are seven layers. The most important is probable layer 3- the network layer. This deals with multiplexing various calls over a single line. It deals with setting up calls across a network.
In fact layer 3 and below are the well known X25 protocols and have been in use for a long time. Above layer 3 the protocols are less well developed.
There was no argument against X.25 being selected for wide area networks. The PTTs had developed it and it was already well defined. All that was needed was to re draft it in ISO language.
It is worth noting that there are alternatives to X25 for local area networks and for ISDN. However the beauty of the layer structure is that above a certain level the protocols do not change regardless of the underlying physical network.
So- X25 deals with layer 1- the physical layer which is all about plugs and sockets and voltages. Layer 2 is all about the connection between a computer and the network and talks about check sums, retransmission and so on. And layer 3 has already been mentioned as being responsible for making connections across the network to a remote computer.
This is layer 5 and a very curious beast. It presents a certain quality of service to the higher layers. Above this layer you should not have to worry about the technology being used.
I group these together and any application demands a certain flavour of presentation and session. Session is all to do with setting up and maintaining a connection while presentation talks about character codes and the like.
The PTTs, in order to get some use out of their networks, defined some rather revolting interactive protocols known as triple X as they consist of three protocols. These defy the ISO layer structure. They have not got an ISO standard but sit there well used and hated. Hopefully they will in time be superseded by a protocol known as Virtual Terminal Protocol which will be an ISO standards and will be well defined and well produced. Regretfully it is currently only partly defined. None the less we can live with triple X for a bit yet.
Mail is popular. In fact very popular. So popular that the PTTs are tempted to run mail services such as TELECOM GOLD. They see even more revenue if they can make it international. Hence they have defined the so called X400 series of protocols (also known as MHS) which allows their mail boxes to be interconnected.
This time they did the decent thing and the session and presentation layers are the ISO ones and the application layer is going through ISO.
This work is very well advanced and examples now exist both in the PTTs and outside. It is highly likely that the international electronic mail problem is next door to a solution both for public systems and for private ones.
This applications protocol is all but defined and implementation is going ahead. It will allow file access as well as transfer. It is difficult to get excited about this useful and essential protocol.
Now here is an interesting one. It is complicated by the fact that jobs have output and moreover jobs can spawn other jobs elsewhere and, moreover, a job may want to collect input from all over the place.
So it is complex and not yet well defined although very good progress is being made.
Back to the humble terminal. The user wants VT100, IBM 3270 and perhaps a lot more.
Hence this protocol is complex to deal with these requirements. As it is a long way from complete a simple subset has been defined which will deal with fairly simple terminals. So here we have a rather nasty hole in the protocol suit.
As with the wide area there are many companies solving all your local area network problems.
Most if not all the suppliers only solve half the problem. They can string together a load of IBM PCs, They can connect a load of terminals to a computer. Very rarely can they connect all the computers and terminals on a site and offer a wide range of applications. It is a case of buyer beware.
None the less the ISO standards are very well advanced. This work was started by the IEEE in the 802 project. They defined a number of technologies. These were the ethernet, token bus and token ring. These technologies are now well established and are a bit expensive. You can certainly now buy these products for most machines. However there is a serious problem in that they only usually support proprietary protocols and not ISO ones. This means that the supplier can only link together certain machines for certain uses.
For example, there are several ISO network products connecting IBM PCs using proprietary higher layers which can therefore not talk to large IBM machines or anything else. You can get ethernet between DEC machines using DECNET which will certainly not communicate with IBM PCs although they may be on the same network.
Hopefully in time these problems will be solved.
As I have mentioned the ISO protocols are complex. In practice only a subset of a protocol is used. Therefore we have a problem of examples of protocols being unable to inter work because different options have been selected.
To overcome this problem CEN/CENELEC (the European standards organization) and CEPT (the European PTT organization) are getting together to define 'functional standards'. These define exactly how you should use the ISO protocols to ensure inter working.
This work is now very well advanced and will ensure that products from a manufacturer with a plastic label saying 'conforming to functional standard xyz' will inter work with similar equipment.
What is the object of a manufacturer? The answer- to make money. Thus if he makes money by following ISO and producing the goods outlined above then he will do so.
So- if you want all this inter working you will have to ask for it and ensure that the manufacturer will not get so much money out of you unless he does so.
In the academic area this is being done by only purchasing equipment from manufacturers supplying certain communications products. Just now there are a load of non proprietary protocols cooked up by the academics. None the less even IBM is now supplying these products as it is still licking its wounds from loosing many orders before it did so.
If a few academics can get such a response it should be far easier for a much bigger customer base.
The signs are good. Both DEC and IBM are producing ISO products however the routes they are taking are different.
The IBM statement is that they see ISO protocols as a way of interconnecting incompatible network systems such as SNA and DECNET. This statement will certainly keep their current SNA users happy and give the ISO fanatics some hope. My own view is that IBM are feeling their way forward. They have certainly progressed from little interest to having products up to and including presentation in a very short time. I hopefully expect them to come out with more products which will attempt to mirror SNA functions over ISO protocols. Whether this will lead to a graceful migration to a full dependence on ISO even though they may still call it SNA is an interesting speculation.
DEC have been far more forthcoming in that they intend to adopt ISO protocols where ever possible. It does appear as though they intend to migrate DECNET to ISO although this will certainly take some time to do. The strategy appears to be to migrate the lower layers to start with and the upper layers later. The current VAX X.25 product is a very good product and represents a very healthy start. Very soon we expect the mail X.400 product to appear. There are rumours that file transfer is being worked on.
Other manufacturers are also dedicated to ISO. The strongest group are the European ones who are involved with ESPRIT. Here there are a number of exciting projects to further the production of implementations.
In the USA there are a number of exciting developments.
The General Motors MAP project and the Boeing TOPS project are both based on ISO and the work is towards using ISO protocols in manufacturing and office automation respectively.
Recently the COS organization has been set up with several million dollars by manufacturers to foster open systems interconnection and they are working on similar activities as CEN/CENELEC to produce functional standards.
Lastly the USA National Bureau of Standards have a very strong activity.
There is still room to believe or not to believe that ISO will come and will be all pervasive. It has come from being a rank outsider to a possible favorite in a matter of a few years and it is still gathering strength.
Since this conference is all about DEC then one must say that DEC appears to be a very strong supporter of ISO with a useful range of products coming along.
Perhaps my own prejudice shows through. I am certainly dedicated to furthering ISO as much as possible since I have to live in a world of heterogeneous machines and have users who demand to pass information between these machines very freely. In addition my users are more and more international which further complicates things.
RCCC has identified a requirement for 'dial back' equipment. The principle requirement is to allow dial up charged to be met directly rather than being recharged by users. A second possible requirement is to improve security by controlling who can use dial in facilities.
Equipment has been identified from Dowty Information Systems (0635 33009). There may be other suppliers.
The equipment is located between the asynchronous modems and the computer or PAD interfaces.
It consists of a controller which has two terminal interfaces- one for control and one for logging. The controller can control between 1 and 54 asynchronous channels. Each channel needs a 'channel card'. The controller can accommodate up to 10 channel cards and extra cards are in subsidiary crates. The equipment may be accommodated in 19 inch racks.
The modems must be Hayes compatible. The Hayes compatibility is required to allow the equipment to dial out. The Laboratory has no such modems.
Controller 5850
Channel card 229 each
Modems 758 each - these operate at all
required speeds automatically.
Total cost for 16 line system 21642
The controller contains a table with 'name', 'password', 'access', and 'telephone number'. 1200 entries may be accommodated.
The speed of the terminal is determined on contact and the controller asks for the users 'password'. If this is recognized the 'access' field determines what the user can do. This will allow the user to access the remote system directly or will dial him back. On dial back the controller will wait 15 seconds for the call to clear down and will then dial the stored number. On the call being re made the modem re selects the speed and the connection to the remote equipment is made.
The 'name' entry is used only for logging.
The 'channel cards' appear to operate at 2400 bps and the modem supplied by the company convert any line speed to 2400 bps. To prevent over-run the modem has a fairly large buffer and operates XON/XOFF protocol.
The 'logging' interface on the controller logs every call and this could be used for accounting purposes.
The tables in the controller allow considerable control in that users may be restricted to make calls at certain times and defines the conditions of access. Thus it would be possible to restrict access to certain people and some may be able to make direct connections and others would be via dial back.
If dial up charges are to be returned to the user then the output from the log interface would have to be processed. It is not clear whether an electronic connection to the controller to collect figures in a more automatic way is possible.
There would be some overheads in setting up and updating the tables. This appears to be a manual task and it is unclear whether a more automatic method would be possible.
The equipment would not work reliably into some computers which did not support XON/XOFF. It would work into a JNT PAD.
Both MF and pulse dialling may be used.
The modems offered are more advanced than the ones currently installed as all the required speeds can be supported. At the time the current modems were purchased there was no such equipment.
It is unclear whether these modems would have problems as the current ones do. However these are now being understood.
The currently installed modems could be relocated for use with terminals as they are all compatible due to the adherence to standards.
The question to be asked is whether the facilities are worth 21000 pounds and can the effort to install and look after the equipment be found?
It has not been found easy to fund the Association from the European Community and it is now proposed to levy a subscription based (to some extent) on a countries GNP. This will raise 256,000 ECUs. This will be for running the secretariat and includes 50,000 ECUs for the production of a newsletter.
It is expected that the European Commission may fund a number of projects possible in conjunction with Eureka.
Associated members will not (on current plans) attract a subscription.
Countries will certainly be full members but it is unclear whether where will be any others. In particular the Commission itself, CERN and EARN will probably be associate members. None the less these organizations are hoped to have a close association with RARE. Manufacturers will also have a relationship.
The Association is being set up under Belgium law and the statutes were considered. These were generalized, to some extent, to allow some flexibility in the affairs of the Association. It is likely to be set up in mid summer.
The letter from David Lord inviting Peter Linington to a meeting to consider cooperation was given sympathetically attention. It is regretted that Peter is unable to meet on the 3rd. or 4th. The RARE Association still has a mixed view of EARN and several members would like to see cooperation between them.
This will be held in Copenhagen 26-28 May 1986 at Domus Technica, Gammel Strand 18 Copenhagen K. There will be places for some 150 and it was thought that if it was much larger the workshop atmosphere may be lost.
As with the last workshop, there will be quotas for each country based on population ranging up to 12 places for the larger countries. Places have also been reserved for the Commission, CEPT and CERN. A certain number of places will be at the discretion of the organizers.
There will be more emphasis on technology and less on politics this time.
The first part of the workshop will be concerned with reviewing progress with the Association and with the 8 areas of special interest. Progress with other related projects will also be reviewed such as ERN, Functional standards, ROSE, THORN, Amego and so on.
The middle part of the Workshop will concentrate on the major technical areas such as X.400. FTAM, MAP, gateways and transition problems will also get some attention.
Lastly there will be some presentations on more speculative or long term topics such as ISDN and satellite data transfer.
Financially the Networkshop is aimed at making a slight profit.
One or two draft proposals were considered and there was some discussion on the form that such proposals take and how they should be progressed.
Like EARN, RARE has a small executive which reports to the main representatives. The temporary secretariat, generously donated by Holland, organizes the meetings and the business of the Association. The secretariat is hired from James Martin Associates and they play no part in the political and technical work of the Association.
Present: B Davies (Chairman) J Hutton HEP B Jones Admin C Challenor G&R R E Thomas Informatics W Pulford SNS C Reasons Laser Division P Linington JNT P Bryant (Secretary) Apologies: N Lawrence M Jonson A Bryden will not be attending in future.
C Challenor was not present.
M2.4.4 It is unclear what sort of summary document the meeting required. The opinion was that a strategy paper was needed. It was agreed that such a paper would have to be produced and agreed by the Central Computing Division before being presented to RCCC. A strategy paper will be produced by the Division. Action: P Bryant
M3.3.2 The meeting with BT has not yet been arranged but there was to be a presentation later in the year which J Hutton would attend and may provide the required information. Action continues.
M3.5.1 No project specification will be provided and the item is discussed in section 8.
M4.2.2 The document describing the network had been found to be far more complex than expected. Action continues.
M4.2.3 The paper on Scientific Admins. requirements was tabled as P18.
M4.3.1 Telex is reported in section 3.
M4.4.1 The PAD performance is reported in section 4.
M4.4.2 A letter has been written to CAMTEC and the reply is P15.
M4.6.1 The EARN user documentation has been circulated.
M4.7.1 A paper on 'dial back' is provided as P17.
It is understood that the tender exercise for a TELEX system had been issued and there had been two responses and the one from SYSTEL had been the only reasonable response. It is also understood that it was recommended that the SYSTEL product should be purchased but its current exact status is unknown. P Bryant had seen the response to tender and the request for enhancements to give a good interface to Grey Book mail was virtually absent. It was unclear what proposals there were for this requirement.
J Hutton was disappointed that the views of RCCC appear to have been ignored.
It is understood that a TELEX service using the product is now being provided and some sort of service appears to be in place.
P Bryant reported that Central Computing Division had no objections to the use of TELEX via PSS. In fact the service was outside the control of the Division as it was mounted on a machine under the control of Informatics Division (Grey Book interface) and the PSS gateway (JANET responsibility). P Linington pointed out that the article in the current issue of Network News defined the charges which would accrue to anyone using TELEX and this was not an indication that CCD or JANET supported or encouraged such use.
It was pointed out that the support of the TELEX-Grey Book interface on the GEC machine could not be guaranteed especially in the long term as the machines were likely to be phased out and also the level of effort to support the software was small and decreasing.
Some members said that they had reported service failures but had soon grown tired due to the high failure level. It still appears that the failures are in the IBM - switch link as failures to other machines did not seem to attract complaint.
P Bryant reported the problems that had been encountered in bringing the 4705 into service. A number of modifications had been obtained from COMPRO and there seems to have been some incompetence in supplying the division with software with all the needed updates. There had been problems with the GEC switch which were traced to a requirement for new microcode. The problems have been uncovered by CCD as the use of the system had been very different to UMRCC and READING. The interfaces were being expected to run at a high speed of 48K and were required to have a high throughput. However, recent experience seems to suggest that there were now only minor problems but this is stated with little confidence.
J Hutton asked who was in charge of the parameters used for setting up PADs. He had requested that the setting up to be changed to make some calls easier and the changes had not been made. It also appeared that the IBM was not setting parameters in a satisfactory way. In fact B Day is in charge of PADs.
It was unclear why this problem had not been brought to NDMs attention. It was noted that NDM was a meeting for CCD and 'friends'. It was agreed to open the meeting to the whole site as a forum for discussing technical communications issues. Action: P Bryant
[Secretaries note: It has been found that the IBM does not set its PAD parameters in a satisfactory way but this had not been corrected as no complaints had been received by P Girard. Thus the IBM was operating as specified in the Green Book which was no longer appropriate. This will be corrected.]
In response to an inquiry from Informatics Division it was stated that it is possible to undertake binary transfers by selecting the CONTROLCHARS option.
The meeting expressed concern that there was no 'easy guide' to the use of PADs with details of how popular machines should be accessed. Is there an action?
P Bryant reported that migration to NRS was progressing and a complete transition could be expected later in the year. P Blanshard had been in charge of coordinating the work but since he had left and the work was progressing and there was no effort to coordinate the work it had been decided in NDM to merely monitor progress.
All machines would have some sort of NRS capability in the next two months but in many cases there were teething problems.
Concern was expressed as to how the tables in various machines should be maintained. It was noted that since computing activities and control of computers was now spread very widely over the site strict control was difficult as various managers would have various ideas of how to control their machines. Computing Division was holding a complete NRS table in the IBM and this, or subsets of it, could be extracted for use in other machines but the mechanisms for extraction would be the responsibility of computer managers. It was unclear how satisfactory this would be.
It was unclear what was required but P Linington intended to produce a paper on a sites responsibility for NRS which would be circulated. Action: P Linington
P Bryant reported that about 40 sites had joined EARN and there would be a meeting for members on 12 February. The mail gateway was not yet complete as RSCS had required a lot of support of late. An information service will be offered shortly and this is being developed by a student.
P Bryant reported that some modification to the slow speed dial up in the telephone exchange had been made but it was unclear whether there were still any problems.
Changes had also been made to the high speed dial up and J Hutton reported that these had made the service useless. He regretted that this had been done.
In discussion it was established that the slow speed dial up was a service but that high speed was not and until it was no service was guaranteed and the emphasis was on perfecting the service rather than maintaining a service. However quite extensive tests had shown the equipment working both before and after J Hutton's experience. P Bryant is now taking a personal interest in the equipment.
Details of a possible 'dial back' service are in P17. It was decided not to proceed as the cost was quite high and that the equipment would require effort to maintain tables and to produce accounts. In addition acquiring new dial up modems when there were still unresolved problems with the current ones was an invitation to sink further resources unproductively.
SNS reported that they had a 'dial back' service of their own which relied on the VAX acting as a PAD. Members were invited to try it.
P Bryant put foreword the proposal that he should get a student to mount a mail directory service based on the GEC mail directory service and one of the existing personnel data bases. There was considerable activity in personnel data bases on site and it was more appropriate to take a pragmatic approach rather than expend effort on a fully worked out proposal that may be over taken by events quickly.
There was worry about the support of such a system but the reality is that because of the lack of effort it was ether such a pragmatic approach or nothing at all.
P Bryant added that he would aim to provide a GEC like name service on the IBM based on the Columbia mailer and move the service to the IBM in view of the short life expectancy of the GEC.
It was agreed to support P Bryant's proposal.
Wednesday 25 June 1986 at 14.00 in R31 conference room.
J Hutton initiated a discussion on 'hacking'. It was clear that this was a sensitive and difficult topic and the exact steps being taken should not be publicized. It was agreed that sites seeking information to catch hackers should be given any relevant help but that security on other sites could not be Rutherford's business. It was agreed that B Davies should contact other site computer managers with a view to increasing security. Action: B Davies
M2.4.4 Produce network strategy paper. P Bryant M3.3.2 Arrange meeting with IPSS. J Hutton M4.2.2 Provide document describing network. B Davies M5.2.1 Provide network strategy paper when complete P Bryant M5.4.1 Open NDM to whole site P Bryant M5.4.2 Produce PAD guide ? M5.5.1 Provide 'NRS site responsibility paper' P Linington M5.10.1 Deal with security. B Davies
Present: P. Lingington JNT (Chair)
B. Gilmore ERCC
J. Craigie JNT
N. Davies AUCC
M. Guy Cambridge
W. Black NPNCG
A. Dransfield LNT
D. Jackson ULCC
P. Bryant RL (Secretary)
Apologies:None
Resignations: J. Larmouth. John has now taken up other duties which will prevent him taking part in the Transition Groups activities. His considerable contribution to the activity is well known and appreciated. His active support will be sorely missed.
The aim of this meeting was to produce a final draft for circulation.
The document will go to the Computer Board as a draft for comment but it is not expected that they will wish to change the technical content but will wish to appreciate the issues. They should approve wider circulation for comment.
It should be possible to receive comments from the wider community before or at Networkshop with full comments by mid May.
2.1 A draft copy of the paper on Country Codes had been circulated and a final version is now expected. Action: S.Kille
2.2 The X400/822 mapping paper was available.
2.3 The exact situation with respect to the use of binary transfers was still unclear and the impression was that it would remain so. The situation with respect to most popular machines was clear but there were smaller communities with requirements and it was considered that these would best be solved on an ad hoc basis. An action is retained to review the situation later.
A reply is still expected from DEC.
Multics requires the transfer of binary strings. Action: All
2.4 P. Bryant reported that no list of members or even chairmen of the CEN/CENELEC/CEPT Functional Standard working groups had been released. If one is produced he will circulate it. Action: P Bryant
P. Bryant agreed to circulate the latest draft of the Y11/Y12 working group and urged members to comment before the final drafting meeting on 12th February. Action: P Bryant
2.5 An updated statement on TELEFILE was expected in mid summer 1986. B Gilmore reported that the company were interested in producing a Yellow Book to X.25(84) converter gateway. The action is retained to review progress. Action: B Gilmore
The complete draft was reviewed. Only areas of contention or where further work is required are noted.
3.1 In principle all combinations of Coloured Book to ISO conversion should be allowed although priority should be given to network service conversion.
3.2 A lively discussion concluded that it was not sensible to expect Grey Book over Blue Book or Red Book over Blue Book to go through a Blue Book FTAM converter with any useful results as these Coloured Book products were to tightly bound together.
3.3 It was unclear as to the relationship between tracking and conversion gateways. It may be that one would develop into the other.
3.4 An index is required. It is quite hard to find where SECTION TWO and SECTION THREE begin.
3.5 Addressing had a lively discussion but this time related to the drafting of the sections and the section is now technically agreed.
3.6 W Black agreed to write to Sir Keith to encourage the appropriate authority to hasten the registration of the community in the ISO-DCC scheme. Action: W.Black
3.7 After a long discussion it was agreed that "context" was related to mail, FTP etc. and "network" referred to PSS, JANET etc. There was a call to make these concepts clear by the use of more examples.
3.8 Consideration of 4.2.1 led to a discussion on the uniqueness of reverse NRS look up with respect to converting gateways. It appeared that a reverse look up may well yield a number of options and the selection of an option would depend on circumstances.
3.9 Three members were concerned that support through gateways for "take job input" and "take job output" was not to be supported. These are in heavy use. In principle they were not needed if JTMP was in place and a migration to JTP was defined. Unfortunately the time scales for the JTMP and JTP products and gateways was very unclear which can be interpreted as meaning that in many cases JTMP would not be adopted at some sites before FTP was phased out. It seemed unlikely that JTP would be available before some FTPs would be ready to be phased out. If recommendations were not produced then ad hoc solutions would be produced and these may be different for different sites and thus involve a lot of wasted effort. It was suggested that some fairly naive scheme based on reserved file names or user-ids would be adequate and there was little point in attempting some elegant (and expensive?) solution.
Mon 3rd March 1986 at 10.30 at Elizabeth House.
Present:- David Lord CERN Chairman Paul Bryant RL Technical Director Herb Budd IBM Paris Alain Auroux IBM Paris Dennis Jennings University College Dublin and NSF Stephano Trunphy Italy Secretary Jean Claude Ippolito Montpellier Treasurer ? Germany Apologies:- Birgitta Carlson
There were none but the minutes of the last full BOD meeting were circulated. It was agreed that it would be useful if minutes were of the executive were produced.
There had been an 'official' opening of EARN in France which had been attended by 'top officials' of government. It had been reported in 'Le Monde'. A video publicising EARN had been produced and this may be used at the Hanover Fair (March 10-19) where the EARN X400 product would be demonstrated.
Germany will not have an EARN opening.
P Bryant reported that the X400 working party would meet for a week at the end of March. The 8 sites were now defined and technical experts identified. He made it clear that an important aspect would be to examine the conformance of the product with the CEN/CENELEC/CEPT Functional Standard. Other aspects to be examined would be reliability, user interface, interworking problems, relationship to other IBM products and so on.
X400 under MVS was not in the project as there was no product. It is assumed that MVS would get X400 services via RFC822 to X400 gateways which would be provided as part of the Heidelberg product. There are problems with JES2 and JES3 with RFC822 and the Technical Group to look at it. A mailing list will be set up for MVS RFC822 problems. Action: P Bryant
P Bryant expressed disappointment at the extent to which RFC822 had been adopted in the community and he was asked to raise this again at the Technical Meeting. Action: P Bryant
There is quite a bit of interest in running X400 over RSCS which should be investigated.
There was a lengthy discussion and the classes of membership were further refined. The principle problem is Associated Membership. In particular there are a number of research divisions in industry which need to communicate with work in Universities. It is clear that any such research must be in the public domain. It does not yet seem that the classes are yet satisfactory. Associated Membership will have to be confirmed once a year.
It appears that no volume bills have yet been raised in the community and it appears that such charging is giving the PTTs problems. It appears that all German leased lines (EARN and all) would attract a volume tariff. It appears that the European railways run a network and this attracts a flat tariff as does NATO. Thus the PTTs are not being even handed.
There was a lot of inconsequential discussion on the PTT's attitude and on their regulations and there were as many views as members. D Jennings considered that the PTT's activities were illegal. There was also a view that statements from the Commission on tariffs across Europe would be useful ammunition and D Jennings will circulate a document on the subject. Action: D Jennings
D Lord will draft a letter to the PTTs which should be sent by each country to its PTT and appropriate Ministries.
There is some talk of RACE wanting to use the EARN network. As yet there is no proposal and the details of the idea is unclear.
Brussels 25 March starting at 9 am and lasting two days.
The discussion on a large EARN meeting was continued.
Last time the idea was to have users as well as suppliers of EARN. It was felt that it would not be possible to run such a meeting as there was little time to organize it and it had to wide a spread of interest.
Another proposal was to run t as a 'World One Net Meeting'. This could be seen as related to the Landwebber meeting and as this and now been organized in Dublin two meetings would be too much. Thus such a meeting may be held in 1987. It was also noted that there were rather a lot of meetings this year. In particular P Bryant outlined the ideas for the RARE meeting.
There was discussion as to whether EARN should sponsor an EDUCOM type organization in Europe. This would appear not to effect RARE interests as it would concentrate on the exploitation of the networks. H Budd and D Jennings will consider producing a proposal for a meeting. Support would be sought from other organizations. Action:H Budd, D Jennings
P Bryant outlined current progress. There was considerable amusement that financial support would now come from members rather than the commission.
There was encouragement to cooperate with RARE and it was felt that the two organizations had a lot to offer each other.
D Jennings outlined the consortium of mainly American organizations who were implementing ISO protocols. He was sad that from a mainly European start the Americans had now grabbed the lead . They were putting $5M into the consortium and the secretariat had a staff of 60. There was a lot of interest in TP4 which he thought was one of the main ISO products. It appeared that X25 may not be the main network substructure in the USA.
It appears that a USA (and possibly other) export licence would be needed fir EARN connection in East Europe. The Board were not happy with this. H Budd advised the Board not to get involved with such connection as the paperwork to achieve anything would be very great.
As this is the first Directors report it is useful to outline the way in which EARN has been developed.
In mid 1983 Rutherford was invited to host a presentation of BITNET by Ira Fuchs from the City University of New York. He described the BITNET network in the USA. BITNET had started from a small group of sites observing that it was possible to connect together a number of IBM computers using the RSCS protocols and provide a file transfer and mail service. The resulting network works on a store and forward basis which prevents useful interactive services but none the less provides very valuable services at a reasonable cost.
From these humble beginnings this network grew to connect a few hundred machines. It was therefore necessary to set up some rules and regulations for the network. For example who can join is carefully defined as are the benefits and responsibilities of a site. Broadly there are 'full members' who can communicate freely and who are generally educational institutes. 'Associate members' are government and commercial organizations who have a need to communicate with a full member usually for some specific purpose. Associate members cannot communicate with each other. This is a simplistic view but in general the idea is that the network is for academic use from academic institutes while trying to allow legitimate academic traffic to non academic organizations.
The development of BITNET had been supported to some extent by IBM.
IBM was keen to see a similar network in Europe as they felt that there was a need for such a facility and that there seemed no reason why the same technology should not be used. Thus the presentation to Rutherford was one of a services across Europe.
It was clear that IBM was offering to fund the international part of such a network and it was also clear that there was considerable support in the academic community.
In the case of the UK it was very clear that the development of such a network would be to some extent in competition with JANET and that this could adversely effect JANETS development by diverting funds and human resources. It could also invite the development of other networks which would likewise be incompatible with JANET.
The solution of the dilemma as to how to take advantage of EARN and yet prevent a separate EARN network in the UK was solved by deciding to provide a gateway between the two networks at the Rutherford Laboratory. IBM generously agreed to fund the development of this gateway by paying for 6 man months of effort.
As with BITNET, EARN has had to produce rules and regulations for the use of the network. This is to stay within the law and also to ensure that the network is used for its intended purpose
It was made clear by BT that if the UK was to be connected to EARN then a special license would be required as the network contravened the BT monopoly on switching.
From the Autumn of 1983 until September 1985 negotiations dragged on. Talks started with BT but it became apparent that with the privatization of BT they were not very sympathetic. As soon as privatization took place BT advised that DTI should be spoken to. So in August of 1984 negotiations restarted from scratch with DTI. It turned out that DTI were heavily overloaded with licensing requests and the EARN one was not high on the list. After a large number of meetings, letters and anniversaries a license was finally obtained in September 1985. Problems were not over as the tariff had to be agreed. BT, on advice from CEPT, imposed a volume tariff based on the IPSS charges. Although this is very high it proved impossible to lower it although negotiations are continuing.
The service started in October 1985 for file transfer. Lines to CERN and UCD Dublin are installed and operate at 9.6K. The Dublin line now carries X25 traffic.
EARN, contrary to popular belief, is not run by IBM but by a Board of Directors. There is one Director from each country. The Directors meet once or twice a year. A small executive committee meet rather more often and deal with most of the day to day business.
The current executive is the Associations officers who are:-
David Lord CERN Chairman and standards Dennis Jennings UCD Dublin Past Chairman Stefano Trumpy CNUCE Pisa Secretary Jean-Claude Ippolito Montpelier Treasurer and promotion Paul Bryant Rutherford Technical coordination Birgitta Carlson Stockholm Hargan Hultsch GMD Darmstadt CEPT liaison Herb Budd IBM Observer and applications Alain Auroux IBM Observer
Minutes of the Board meetings are available on demand.
The Association has been set up under French law.
Finance for the association and the international part of EARN is provided by IBM until the end of 1987 after which funds will have to be provided from elsewhere.
The gateway initially provided for file transfer using a technique developed in the past for moving files to and from the IBM computers at CERN.
It is recognized that this is not very satisfactory from sending files from EARN to JANET and so an IBM exec has now been provided which removes most of the problems. It has the major advantage in ensuring that the success or failure of a transfer is reported to the user and in addition allows files to be 'fetched' from JANET nodes as well as being sent.
A major use of the gateway should be for mail. The mail gateway is now almost complete. This will allow the much freer exchange of information as details of remote file stores and pass words will not be needed. Once the initial service is launched it is intended to enhance the gateway to provide distribution lists and more sophisticated addressing mechanisms.
It is not intended to provide any services apart from file transfer and mail. In fact the only other service that EARN provides is 'messaging' (that is a real time one line message being sent to a logged in user) and it would be fairly difficult to provide this. In particular most nodes in JANET have no comparable facility on which to map it. Interactive service are not provided as the network is based on store and forward techniques which would make response very poor, erratic and unreliable.
EARN has an information service called NETSERV. It is mounted on each international node. It contains a miscellany of services and information files to help the users and aid management of EARN. The files designed to aid users have now been made available via 'Blue Book' file transfer from the IBM computer at Rutherford. The management files have not been made available as these are used for updating various routing tables, for network error reporting and other essential activities.
EARN will shortly provide a name directory for people which is being developed at Paris. It is unclear whether this can be made available from JANET but this will be studied later in the year. There is a conferencing system which EARN hopes to mount called GRAND. Again, how this can be made available is as yet unclear.
Unfortunately a gateway more often than not reduces the 'quality of service' that is seen. This is particularly true of this gateway due to the very different characteristics of the two networks. In addition the effort available to improve matters is very limited. None the less it is hoped that the service will improve as it develops over the months.
About every six months there is an EARN technical meeting where experts form each country discuss the problems of the network. The first was held in the UK, the second in Mainaz near Geneva and the third will be in Paris in a few weeks. Minutes are available on demand.
High on the list of topics is reliability of the network since, unlike an X25 network, it relies on availability of large machines.
Mail, conferencing and migration are other important topics.
As a result of allowing EARN to exist CEPT (the advisory body for the European PTTs) has required EARN to migrate to the use of ISO protocols and to use public networks by the end of 1987.
To this end a project is underway to use the mail service based on the new X400 mail standard. This will link 8 international EARN sites.
The system is being developed by the IBM Heidelberg Scientific Centre and will be demonstrated at the Heidelberg Fair in the spring. IBM has generously funded a six month study to find out is the product can be the basis of the migration. Rutherford Laboratory will be taking part in this work.
It is unclear how EARN will migrate but this should become clear as the project progresses. It is likely that BITNET will also migrate to alternative protocols and an aim will be to ensure that gateways are built between the networks.
There are two problems. The financing of the lines and the long term financing of EARN.
Since the PTTs have levied a fairly high tariff it seems unlikely that the funding for the lines from IBM will be sufficient in the UK. It should be added that although the UK has been first to impose such a tariff the other PTTs show signs of following. It is difficult to estimate how long the money will last as this depends on use but if the current increase in usage continues it will run out after 6 to 9 months of use. The service will then cease. In fact IBM has supplied 21890 pounds to last until the end of 1987. The BT charge for the CERN line is 4766 pounds per year plus 1.1 pence per 1000 characters subject to a minimum of 7960 pounds. The Dublin line is 2472 pounds per year and 0.94 pence per 1000 characters subject to a minimum of 4120 pounds. It should be added that all UK to CERN traffic is volume tariff free. In addition the Dublin line uses X25 and should probably be treated in a different way.
There are a number of options:-
The last is of course the best and these is some indication that the PTTs recognize that the type of traffic carried by EARN is large and of low value and should thus attract a low tariff. Unlike airline traffic which is low volume and high value. Such a tariff would have to be agreed across Europe and the Board of Directors is currently negotiating with CEPT.
The long term financing of EARN is difficult to discuss as the future form of the network is unclear. Should it migrate to use public network then the consumer would pay and the PTTs would levy their normal charges. In fact EARN would more or less cease to exist as a physical network. If it does continue using leased lines then about 15,000 pounds a year would need to be found. That is something like 500 pounds per site. The list of options is similar to those for funding the short fall which were outlined above.
The minutes of this meeting will be submitted to the EARN Board of Directors. This should give them a view of how the UK sees EARN and would like it to develop. It will highlight problems to them.
Each country has to select a Board Member. How this is done is not specified. Thus, if the meeting agrees, this meeting should elect a Board Member. It is suggested that each member should have a vote if any election is needed.
The duties of the Board Member is to supervise the operating of EARN in his country and to represent his country at the Board of Directors meeting. In the UK case it would involve making sure the gateway works and sorting out any problems. It would involve one or two meetings abroad lasting a couple of days. The position is unfunded and should not absorb more than 25% of a persons time. The position of an executive Board Member and the position of Technical Director would not necessarily be vested in any new member.
How the UK organizes itself is up to this meeting. You may like to elect or appoint a steering committee. You may like to decide how often meetings such as this are held. You may like to consider how this group should communicate its views to the Board of Directors.
NETSERV is an information service machine providing general services to EARN users in all countries and some extended services to certain users in the United Kingdom.
Data files of various types and contents a available from this service using the methods detailed below.
The file access methods have been augmented to allow JANET users to use this service in a different way to other users. However, all the facilities of NETSERV will still be available to them.
It is recommended that JANET users access NETSERV's facilities in the special manner detailed below as this reduces the computing overheads on the NETSERV machine.
Two types of user are supported by the NETSERV facility. General users are only allowed to retrieve selected files from the NETSERV file store.
Privileged users are, in addition to the general commands, allowed to send files of their own to the NETSERV facility so that they may be included in NETSERV's file lists and be made available to other NETSERV users.
Only privileged users are permitted to send files for storage in the NETSERV file store. Before files can be stored on the NETSERV file store you should contact the NETSERV controllers at R.A.L. The controllers are :
Tony Burraston AWB @ UKACRL Andrew Wilson ARWR @ UKACRL
and they will provide you with a password and other information that you will need.
Normal users are those users with userid's on the IBM370 at R.A.L. These users may talk to NETSERV directly as detailed in the document
NETSERV HELPFILE using the TELL and GIVE commands on the IBM. The section 'Further Information on the NETSERV facility' explains how to obtain NETSERV HELPFILE.
A list of the files available to JANET based users of the NETSERV facility is contained in the document FILE$.LIST$ . This document can be obtained by using a file transfer program like FTP on the IBM or by using the equivalent program on your own computer.
The information required by your file transfer program is :
remote_file_name = FILE$.LIST$ remote_site = RLIB remote_username = NETSERV
Note. There is no 'E' on the end of NETSERV. No passwords are required.
The FTP process will send you a file containing the names, disk locations and ... descriptions of the files available. Some of these files are located on different disks of the NETSERV system and for these files the disk location may need to be included in the remote_username field as required by your FTP program.
For example. If the file EXAMPLE.FILE was located on NETSERV's 193 disk then the remote_username given to your FTP program would be :
remote_username = NETSERV//193
Note. There are no spaces between the word NETSERV, the '//' or the number 193.
If JANET users wish to store files in the NETSERV file store then they must first inform the NETSERV Controller and obtain a password, see 'Becoming a Privileged user of NETSERV'. Once their password has been obtained JANET users may store files in the NETSERV file store by methods detailed in the document NETSERV HELPFILE. The process is quite simple. A suitable command is inserted ..... into the first line of the file you wish to store and the file can then be transferred to NETSERV by the FTP program on your machine.
As detailed in the document NETSERV HELPFILE, suitable commands are placed on the first line of a text file which can then be sent to NETSERV by the FTP program on your machine.
Further information on the NETSERV facility
A substantial helpfile detailing the commands and facilities available to users of NETSERV can be obtained directly from the NETSERV facility in a number of ways :
Provided that you have a userid on the IBM370 mainframe at R.A.L the help file can be obtained in the following ways : Either entering >>>> MORE
TELL NETSERV HELP
at your keyboard, or similarly
VNET SMSG UKACRL NETSERV HELP
and NETSERV will send you a large document (about 1600 records long), which will appear on your file reader. Alternatively a file containing the word 'HELP' on the first line can be sent to NETSERV using FTP or the R.A.L. GIVE command.
The NETSERV help file can be obtained by requesting a file transfer of the file NETSERV.HELPFILE from the NETSERV file store using FTP or the equivalent file transfer program on your computer. The following conditions are required for the transfer :
remote_site = RLIB remote_file = NETSERV.HELPFILE remote_user = NETSERV//193 >>>> MORE
Note. '/' is the divide sign and no password or account number is required by the program.
It was thought that the name EARN had been registered in the NRS for sending mail between EARN and JANET. Unfortunately this is not true and there are some problems which still have to be resolved. Had EARN been registered it would use JANET DTE 000000000002 and Yellow Book string FTPR.MAIL.
It is unclear whether the name EARN on its own will be registered in this way. It will, however, be registered so that mail can be addressed to Fred@EARN.site when this form of addressing is made available. Until a decision is obtained from the NRS administration users may like to make use of the above information.
It should also be pointed out that the name UK for mail in the EARN to JANET direction has a similar problem in that the use of UK is being debated and that it had not been built into the mail systems of EARN yet.
Dear Sir/Madam,
Electronic mail is a wonderful thing- when it works. In this case it has not. Life is too short to ferret around when it does not so I am sending you electronic mail which I think should have ended up on your site by post reliable means. The mail was generated with a 'reply' facility on a GEC computer.
Please would you try to deliver it- if you can find out who it is for.
It may also contain information useful in enabling mail to be delivered in the future.
Yours sincerely
Paul Bryant.
Dear David,
At our recent telephone conversation I agreed to write down a few notes about the facilities that can be offered by JANET. Your plans to use TELECOM GOLD were brought to the attention of the SERC Communications Committee who were surprised that you were not taking advantage of the extens pb260 EARN Gateway manual 14 Feb 86
pb261 Mail document for GEC 14 Feb 86 pb262 Letter David Jones Teaching Company 14 Feb 86 pb263 Addendum to gateway document 14 Feb 86 pb264 Letter Y11 Y12 on AITAEGS 22 Feb 86 pb265 Document for Carlos 22 Feb 86 pb266 Documents for EARN TECH 22 Feb 86
Dear Sir,
Please supply two 8088 compatible NEC V-20 PD70108 chips which I understand cost $20 each.
Please use VISA card number ... in the name of DR P E BRYANT.
Yours faithfully
P Bryant.
Dear Sir,
We recently bought a MALTRON keyboard from you for evaluation. We are impressed with the layout which confirms your claims. However the device has taken more getting used to than you suggest but about as much as I expected. There are a number of problems and observations.
We use Wordstar and the control functions are grouped in a logical way for QWERT keyboard. The MALTRON keyboard destroys this relationship. There is no answer to this one.
When used with one hand (phone in the other) the keyboard becomes awkward to use. I see no solution to this either.
The keyboard delivered had an American layout. There should be a UK option to avoid the bits of sticky tape.
A coiled connector cable as supplied with the IBM one would be a neater job.
The keyboard is very light and tends to flex a bit and this is not liked by typists who seem to prefer heavy and solid devices. These have less tendency to slip about.
The type of key is very inferior to the keys on other IBM keyboards. They have a 'soggy' feel and typists prefer the click action of the IBM one.
I hope these comments are of interest. I hope the keyboard has a successful future.
Yours sincerely
Paul Bryant.
Dear Dr. Raubold
Last year a number of meetings were held to decide how EARN should migrate to use ISO communications protocols. I expect you are well aware the CEPT is pressing the EARN Board of Directors to use these protocols by the end of 1987. Eventually it was decided that the Heidelberg/GMD X.400 was the best starting point for this migration.
Subsequently the Board of Directors obtained financial support from IBM to run an experiment during 1986 to evaluate the system and to gain experience in its use. Eight international sites have been selected to take part. The experiment will start with a course at IBM Heidelberg during the week of the 7th April.
Since GMD has been deeply involved in the work it would be very useful if a number of your staff could assist at the course. I believe that Dr. Thomas Schuett has already approached then and we have drawn up a provisional agenda including then.
I would be most grateful if you would agree to your staff taking part and ensuring the success of the project. I hope you will forgive me for making arrangements without consulting you until now. We have had considerable difficulty in arranging suitable dates and time is not on our side.
I have taken the liberty of enclosing invitations to your staff and if you are agreeable I would be grateful if you would deliver them.
With best wishes
Paul Bryant. (EARN technical coordinator) PEB@CERNVM
Dear Professor Petrie,
Thank you for your letter concerning the gateway at Rutherford between JANET and EARN. You make a number of very legitimate comments. I would be grateful if you would make the content of this letter available to your meeting.
There was a suggestion that the JANET user groups should be used as a forum for discussing EARN. This was turned down by JNT and hence a user group has been set up and has, in fact, had its first meeting. This is the correct forum for raising EARN matters but is only open to members of EARN plus some observers. None the less, as I am very anxious for EARN to succeed I am happy to reply to your letter.
The first EARN meeting attracted some 40 attendants split equally between computing centre staff and users. This was a surprisingly large number and it was clear that quite a lot of useful work is going on. The points that you mention were also brought up and several useful suggestions were made. The meeting was an opportunity for launching the mail service and the documentation facility.
In summary I am unhappy with the state of the gateway. The operation is funded entirely by Rutherford Laboratory at no cost to the community. IBM fund the cost of the lines to some extent. We used to have a network support specialist but he resigned and it has not been possible to replace him at the salaries we offer. Thus not only is there little EARN user support but there is no user network support at all. There have been a large number of unforeseen problems with BT, DTI and the NRS which have caused seemingly unending delays which made me wary on predicting when services would be offered and the form they would take. Lastly the minimal manpower has resulted in documentation, consultation with users and other desirable activities to be severely limited. Thus there is clearly substance in your observations.
You make the point that there has been insufficient publicity for the service and that publicity has not always gone to the most appropriate people. I would not disagree with this.
The following publicity has taken place. The last two issues of Network News have contained details of EARN and have contained an invitation for anyone to contact me for further details. This resulted in about 20 inquiries all of which were followed up with advice on how to obtain service. A letter was sent to all computing centre directors which were obtained from JNT lists. This also contained posters. This resulted in 40 to 50 sites joining EARN. This represents the majority of the university sites. I have to admit that only a few polytechnics were approached as it was unclear whether they were entitled to join EARN and in any case few had JANET connections. Now this position is clear they will also be circulated if I can find a suitable distribution list.
I am very unclear what other publicity would have been appropriate. I was certainly in no position to write to all university departments as this would have involved considerable expense in postage and considerable research to obtain such a list. If there is such a list I would like to know about it but could only really make use of it if is in form of labels as I do not have the manpower to address envelopes. I may well have been criticized if I had approached departments rather than computing centres as from previous exercises it has been clear that some sites are unhappy with approaches of this sort not being through the centres. I could also have approached major groups such as HEP. Here again I have no lists of such organizations. I should also point out that it has been made clear to me that the use of JNT meetings for publicity is inappropriate as EARN is not the business of JNT and vice versa. I should also add that I was unable to give an EARN presentation to Networkshop presumably as there was insufficient time in a heavy schedule. I am very happy to use other publicity means as long as these are not expensive or time consuming as these commodities are in very short supply. I might add that the reported inability of some computing centres to make sites aware of the EARN service is disappointing but I feel this is an issue for those sites and I should not attempt to repair these reported deficiencies. You are right to criticize the non use of the JANET NEWS facility although I am not at all sure that much use of this is made by non computer experts any more than Network News is apparently read by them- and why should they as they are not primarily interested in the subject.
I am afraid that it was a decision of the EARN Board of Directors to insist on 'sites' joining EARN. I was unhappy with decision but the board felt that UK sites should not be in any privileged position with respect to those elsewhere. The recent EARN user group meeting have asked me to raise this again which I will do. To some extent it is now of academic interest as there are few sites who have not joined.
I agree that support services are minimal if not absent. We have now produced a computer documentation distribution service which was launched a couple of weeks ago. Documentation for this is just being issued. This documentation has been delayed for reasons given below. When Rutherford agreed to host the gateway they intended to use a percentage of our normal network support staff to provide help to users. This is still the case but the unfortunate fact is that our one network support expert left us sometime ago and it has been found impossible to recruit another willing and able to do the job. This is a constant cause of concern to us and we are desperate for such a staff member to support our other network users. In the mean time the burden of support is falling on myself who is certainly not a technical expert. In addition I have many other tasks which take me away from the laboratory. The recent user group meeting felt that perhaps computing centres could take some interest in the network and this looks like a profitable move although I am unclear as to how it could be organized.
I mentioned the delay in documentation. You are no doubt aware that we offer a file transfer service and a mail service. The mail service has taken far longer to come to fruition as the person working on it also looks after other aspects of our activities and instead of spending 90% of his time on the gateway it has been more like 10% due to the high level of activity maintaining other services. IBM are supporting the production of the gateway and we had hoped to recruit staff to undertake the work or rather release existing staff. This has not proved possible, again this is a result of our poor salaries and the county wide famine of IBM networking experts. We have also encountered a number of logical and political problems which caused the recent delays. We found that we could not use the NRS in the way we expected. Also we found that we could not use the names we wanted to within EARN. These problems required some changes to code and documentation which are now all complete and thus the documentation is being released now.
I am afraid that the total network staff which have to develop and support both EARN, our Coloured Book code and IBM style network activities is two people. In addition the management effort we have to spare on EARN is about 25% of myself which covers all documentation and support activities as well as liaison with the rest of EARN.
I would love to have more effort to spend on this activity but it must be remembered that it is a gift from SERC and there is no other funding. I have made some informal approached to JNT and users and it is fairly clear that funding is unlikely from those quarters. I am also in contact with IBM who may be willing to add support. However, even if such support is provided it will not be easy to find relevant staff as I already have vacancies waiting for a suitable candidate.
On a more positive note the file transfer service is not very reliable and stable. Documentation is now stable and available both in the form of a document and access to a large data base of various pieces in information. Mail exists but is shakey but I am confident that this will become stable soon.
I hope these comments have clarified the situation. I am always happy to consider any constructive suggestions for improving the service.
As a postscript you may be interested to know that for the first 3 months (October to December) the traffic rose from 7 to 12 megabytes a month. In January it was 43 megabytes. If this trend continues these will soon be no network as we will have run out of money to pay the volume charges imposed by BT. I would very much welcome suggestions on how to finance the service. An approach to BT failed. Rutherford feel that the resources they are putting in is a large contribution and they see no reason why they should support the network further. JNT have not wished to contribute. It is thought that an approach to the Computer Board would not be successful as they would require to know what the network is being used for and by who. Such information is not collected due to the lack of resources here and the nature of the network makes analysis of use difficult. I am reluctant to consider charging as this would involve the charges to include the costs of producing charging code and manpower to administer it.
Any suggestions for solving our financial and manpower problems would be most welcome.
Thank you for your interest.
Yours sincerely
Paul Bryant. (UK EARN Director)
The IBM 6150 is a personal computer based on a Reduced Instruction Set Computer (RISC) principles.
The machine uses UNIX system 5 with some extensions from Berkeley 4.2. This is called AIX. Memory is from 1Mbyte to 4Mbytes but this is extendible as higher density memory chips become available as the machine has a large address space.
Up to 3 40Mbyte or 70Mbyte Winchester discs and up to 2 floppy discs may be connected. These limitations are temporary as IBM or other suppliers are able to develop equipment with alternative specifications.
8 IBM PC compatible extension slots are available which will allow a wide range of hardware to be connected at an early date although software will be needed to exploit this hardware.
A floating point unit is available.
UNIX has had a number of additions which include:-
Currently about 30 software items are offered which include FORTRAN, PASCAL, BASIC, SQL plus many applications.
3270 access is provided.
Performance figures are not available but it appears to be rather better than the 68000 systems such as SUN or Whitechapel.
In fact the computer is aimed at the university UNIX community and must be seen as a competitor to the SUN.
The IBM 6150 appears to be an attractive machine and may be the vehicle which will popularize UNIX for other than computer scientists. The computer will be aggressively marketed within the universities. Like the PC it is expected that a very large number of identical machines will encourage IBM and other suppliers to supply software well tuned to the hardware and in large quantities and high quality. This is in contrast to the current UNIX scene where there are many different machines of different hardware architecture which means that software may only operate on a subset of these machines. Although UNIX has portability as an objective it is clear that portation is difficult especially with respect to the driving of devices and in particular driving displays.
Before the IBM PC arrived there were systems using PC DOS like operating systems with a variety of hardware and these systems were popular although porting software was usually non trivial. Disc formats differed, the details of displays differed and bus structures differed. The IBM PC 'standard' transformed this by providing a standard for the hardware and operating system which suppliers could work to.
The expectation is that what the PC did for the PC DOS world the IBM 6150 will do for UNIX.
The recent GLM meeting considered that a proposal for the support of the machine by Central Computing Division should be drawn up.
A presentation is being arranged. This should reveal the level of interest in the Laboratory. At this time the Division must have decided whether to support the machine, assuming sufficient interest. At the presentation the Division should announce its intention to support the computer and present its program of work.
It is proposed that the IBM 6150 should be supported in the same way as the IBM PC. That is, the machines should be purchased, installed and supported. There should be little software development but software and hardware from IBM and other suppliers should be investigated and recommended as required by the customers.
It is anticipated that as with the PC the effort required will rise from 1/2 man per year to 1 man per year of good quality labour over a period of between 18 months and 2 years. The exact time scale will depend on the interest shown by the Laboratory and assumes that take up of machines will be similar to the take up of the PC.
The good quality manpower should be obtained by augmenting the current PC support with one person who will undertake many of the duties concerned with the PC.
An IBM 6150 should be ordered as soon as possible and this should be delivered in the Autumn.
The undiscounted price will be 15000 pounds upwards depending on optional equipment. A detailed costing will be produced when further discussions have been held with IBM.
It is likely that as with the PC it will be useful to buy interesting hardware and software for evaluation as they become available and as the needs of the users become apparent.
Dear Alan,
I regret that I am unable to attend the forthcoming OSI/-/2 meeting which is particularly disappointing with the discussion we are having with respect to ITAEGS and the working groups.
Further to our meeting last Friday 14 March and my paper on the problems of Y/11 Y/12 I would like to put forward my view as to how the situation could be improved.
The basic problem appears to be that the project has weak technical coordination. Working groups are to a large extent working in isolation under very loose guide lines as to what sort of documents they should be producing, how they resolve technical issues with other groups and how the quality of their work is assessed.
I would suggest that ITAEGS should become a formal committee for the technical management of the project. The terms of reference could be:-
The membership of ITAEGS should be:-
It is important that the meeting is properly conducted with calling notices, agendas and minutes. Currently the membership of ITAEGS is unclear and very few chairmen attend. To some extent this may be because the only way of knowing when the next meeting is is by attending the last one. In addition there is no way of knowing whether the agenda contains items of interest to the attendee.
The first task of ITAEGS should be to review M-IT-02. A cursory glance at this document reveals it to be inconsistent. Further thought is needed on whether the various work items are sensible entities and whether they are properly defined. For example, a glance at the entries for T31x and Y/11 Y/12 which both use X.25 reveals quite different text and diagrams. In addition it is far from certain that the structuring is sensible. For example, Y/11 Y/12 is aiming to produce the best possible service with the X.3, X.28 and X.29 recommendations but also has to include X.25 in the functional standard - should it merely state the absolute minimum for providing a service or should it include X.25 facilities which are desirable but which may be fairly irrelevant to the X.3, X.28 and X.29 provisions? Although I site examples from Y/11 Y/12 I feel sure that other working groups have similar problems.
The second task ITAEGS could undertake is to resolve problems which various group are having. Thus it is vital for the chairmen to attend as the problems are likely to be common to groups and solutions should be applied to all relevant groups.
The third task could be to review the draft ENVs. This would give the ENV an early review before being presented to a wider audience. It would give an opportunity to ensure that the document was of a good quality and met the terms of reference of the group. The aim would be to decrease the possibility of negative votes.
A major problem appears to be the lack of resources allocated to the central administration of the project. It was clear from the last ITAEGS meeting I attended that the resources comprises the project leader with a small amount of secretarial support. There was not enough resources to circulate lists of the members of the working groups and committees. Also, I suspect that the lack of calling notices, agendas and meeting organization (that we have enjoyed in abundance from BSI) is a symptom of the lack of resources.
Finally, I believe that the project is important and that we should do all we can to make it 'work'.
I regret that I did not attend the last ITAEGS meeting due to pressure of work and the uncertainly of whether there were items of interest to us on the agenda.
Yours sincerely
Paul Bryant.
The Heidelberg technical meeting has now been set up for 7 April to 11 April.
This meeting is being organized by Dr. Thomas Schuett of IBM Heidelberg. Speakers are from IBM and from GMD and are:- T Schuett, G Mueller, B Schulz, D Kropp, W Racke, W Schulz (IBM) L Eckstein, H Ehmke, U Viebeg (GMD)
The EARN participants are:-
H Goossens (Holland), O Martin (CERN), P M Girard (UK), M Auffret (France), H Bergh-Hoff (Norway), J Noland (Ireland), G Pitteloud (Switzerland). I have yet to hear from Italy.
Equipment is being installed at the relevant sites but the exact position is not available.
There seem no further problems in starting the evaluation immediately after the meeting.
At the meeting one afternoon will be set aside for planning. It is anticipated that each site will take responsibility for some particular aspect of the evaluation. For example, reliability, compliance with standards, user interface and so on. Reports will be supplied from the sites at two week intervals. Initially the project will be progressed using the existing EARN mail services. At a later date the X.400 service itself will be used.
The agenda for the meeting is available on request.
The next EARN technical meeting will be held on April 21/22 in Paris.
The meeting is being organized by Dominique Pinse of IBM. The meetings will be at the IBM Scientific Centre which will reduce costs and organization effort. In future I believe that it would be better to use similar facilities as meetings in hotels present financial problems and are far less easy and flexible to organize. It is difficult to predict the important subjects. However, mail (particularly on MVS), gateways, NETSERV and down time will certainly be re-visited. Invitations have been issued to all Board members and participants at the previous meeting. I gather that this has regretfully omitted a number of people for which I apologize and promise to try better next time. Anybody wishing to attend or knowing of others who want to please contact me as soon as possible. A representative from Joiners Associates will be coming and we should have an interesting discussion on the VAX computers.
At long last mail is now available! There are still several developments to be made which are essential for moving from source routing to domain addressing in line with current thinking. However these are not essential for the provision of a service. A problem is the upgrading of all the mailers within EARN to recognize the relevant name for the gateway.
There is always a problem in bringing up a new service as to whether to offer it as soon as possible or to wait until it is fully developed. We decided to offer it as soon as possible and thus attracted a lot of problems in the first week or two of use. Most of these are now solved and after a couple of weeks of use the traffic has risen to 70 messages a day which shows some use is being made even though documentation has only been issued widely in the last day or so.
The EARN/JANET gateway manual is now in issue 3 and is now stable. There will be an update in 6 months when domain addressing is introduced and to correct any errors. A short guide will be produced for the experienced user on no more than one page. The manual is being put in NETSERV.
A user meeting was held recently. This was lively. A principle problem is support. Unfortunately our network support officer resigned some time ago and we have been unable to recruit a new one. Thus, not only has EARN got no user support but all our network activities have none. It is hoped that other sites will be able to give some support. The publicity for EARN was criticized as not being wide enough. In fact all computing centre directors were invited to apply for their sites to join EARN. It was felt that at many sites directors did not represent there populations and had ignored EARN and not kept departments informed. While this is a legitimate point it would have been difficult and expensive to contact all university departments. It would have cost a lot and there is no definitive list of such departments. There was discussion of membership and they wanted membership to be on departmental basis. I am unable to support that on administration overheads grounds. In a way it is becoming of academic interest as 67 sites have joined EARN which is the vast majority of sites.
The financing of EARN is a real problem as the service will close mid year when funds for the line run out. I am approaching various bodies but have so far had no success but am hopeful.
The traffic between EARN and JANET is as follows:-
Total (Mbytes) Charged (Mbytes)
October 12 6
November 80 10
December 52 15
January 231 43
February 91 36
The high January figure was caused by the failure of the non EARN connection between Rutherford and EARN. This line carries about 300Mbytes per month.
If the EARN traffic were to rise to 300Mbytes per month of charged traffic then this would attract a volume charge of 36,000 pounds. Current funds allow for 26Mbytes per month.
There is advantage is moving traffic between the two lines but this has not been possible in general as the Swiss PTT has not set a tariff and CERN are unhappy to allow any such moves until it is fixed.
Dear Mrs. van Heugten,
I have just re read your invitation to speak at SEAS and noticed that you require an abstract and CV. I am sorry not to have replied earlier.
The title of the talk is "The Current State and Future Development of the European Academic Research Network".
The abstract is:-
EARN (European Academic Research Network) is a large international network based on the use of the IBM NJE protocols. The network is used for academic and research pursuits and carries no commercial traffic. The European branch of this network was started in 1983 and has now grown to some 300 machines. There have been a number of problems caused by the international nature of the network and also due to the tariffs imposed by the PTTs. The network has to migrate to use public networks and use ISO protocols by the end of 1987 and this has caused EARN to undertake a migration project.
Here is my CV:-
Paul Bryant drifted into computing after deciding that a career in theoretical physics was beyond his humble brain. He has worked since the early 1960s in Government research establishments on a wide range of computers. He is now based at the Rutherford Laboratory which is part of the UK Science and Engineering Research Council. Since the mid 1970s he has devoted most of his time to networks having been very active in the setting up of the UK academic network called JANET. He is now working on a variety of projects both in the local and wide area fields and also spends over much of his time on standards activities. He is the UK EARN representative as the Rutherford Laboratory operated a gateway between EARN and JANET.
Feel free to prune and modify the above as you see fit. I hope it is what you want.
By the way, I have filled in no forms to attend the conference I hope this is all right. Also, I am making my own accommodation arrangements. I will be looking to you to pay my air travel.
Yours sincerely
Paul Bryant
Complete BOD. Only 5 IBMers, it appears that the attraction is declining. Chris Wilkinson from the commission. Hubner from CEPT did not unfortunately attend.
Arrived too late as per usual.
A request was made for a budgetry report even though only about 5000$ had been spent. The only expenditure had been on travel for executive meetings and attendance at a number of other meetings. Currently there is 50,000$ in the kitty.
USA 781 nodes EARN 344 nodes (excluding 182 in UK) NETNORTH 86 nodes TOTAL 1487 nodes
United Nations in Geneva have applied to join and it was agreed that only certain of their organizations could participate:-
environmental development research organizations UN university world health organization word met office.
BITNET has agreed to give them class D status.
There was an incredibly boring discussion which resulted in the setting up of a membership committee under Birgitta Carlson with Kees Neggers and Frode Greisen to refine the charter document.
There is a request from Hungary to join. Herb Budd stated that there would be an incredible amount of bureaucracies to go through- such as the PTT lines, CoCom etc and the regulations of all these bodies were even then contradictory. It was clear that he was hinting that EARN should avoid the issue. However the BOD generally felt that on academic terms such links were desirable. It was suggested to start a test case. It was unclear what decision was taken, if any.
The old chestnut came out for a new airing. Seems Italy is the latest problem with a suggestion of 8 pence per 1000 characters.
Otherwise the situation is chaos. Of most interest was Spain where one line (Barcelona - Rome) is charges and the other (Madrid - Paris) is not.
Was agreed to ask Hargan Hultch and Michael Hebgen to liaison with CEPT.
My view is that EARN is asking for what it is getting by not having a closer liaison with CEPT.
It was also agreed to produce a paper on the tariffs and their implications.
A Code of Conduct paper has been produced. Discussion centred on whether it had legal status. The answer is clearly no- but it does give the BOD ammunition for expelling ill behaved sites.
A curious organization which is very closely bound with the universities. It was agreed to allow their non commercial part to joins but not as a full member. Interesting- by pushing data through the UK the NL PTT can increase the income of BT! By this means the PTTs can bankrupt EARN by their own unaided efforts!
It is now intended to invite observers from CEPT etc to view the results of the experiment as an indication of EARNs intention to migrate to ISO protocols.
The experiment is now gaining a wider base as the Italians are intent on experiment with the IBM OSNI products and some experiments with VAX X.400 are planned.
It may be that EARN should move to use X.400 over RSCS at an early date as the recommended mail system for the network instead of RFC822. This will be talked about at the Paris technical meeting.
I can not remember a BOD meeting where this topic has not been re-opened to little effect. Again the legal status of the document was questioned with no results. Regretfully the Germans had shown it to their legal departments in each universities with counter productive results. As with most of these documents- they are drawn up with the intent of impressing on members their privileges and duties with the threat of expulsion for non compliance. But in general the document should be put in a deep drawer.
Will be in Berlin on 27/29 October. It will be for 300 people and a Board meeting will also be held. Arrival will be mid day 27th and departure mid day 28th. Michael Hebgen of Heidelberg will chair an organizing committee with Jean Claude Ippolito of Montpellier and Herb Budd of IBM Paris. The program is as yet unclear but it will be aimed at popularizing EARN by talks about its use by various groups and by talks about its development.
The cost, apart from travel, will be born by IBM.
There was a discussion concerning relations with RARE. There was a lot of muddled thinking and some took the view that the aims of the organizations were identical while others regarded EARN as a network provider and RARE as a coordinating and planning body. None the less there was agreement that there should be close ties with RARE and that EARN should seek some associate status within RARE as suggested in a document concerning RARE membership.
There was discussion concerning COSINE. No one could explain clearly what it was about and what it had to do with academic networking. The confusion included the RARE representatives who are also on the BOD.
There are strong indications that BITNET intends to migrate to TCP/IP protocols. This will bring them in line with ARPA and the two networks will combine. From TCP/IP they intend to migrate to TP4.
As yet there is no pressure for EARN to follow this path. In fact, quite the reverse. EARN so far has stated its intention to follow the X.25 path.
The conclusion must therefore be that the networks will part company and that a gateway will be needed between the two communities.
The USA arguments appear to be based on the pathetic quality of the public X.25 networks and the heavy weight of the X.25 stack of protocols. In particular the argument goes that TCP/IP is far better suited to the high speed networks the USA is planning (64K to 2M). Undoubtedly the argument is strengthened by the abundance of TCP/IP implementations available and the lack of the alternatives.
It is now appearing that many European sites are using TCP/IP on their local area networks. UCD Dublin and Informatics at Rutherford are two good examples. This leads to the inevitable conclusion that good gateways are needed between the two camps.
Dear Jan,
Some time ago I raised the difficulty that the work of Y/11 Y/12 Working Group could be sterile if the CEPT standards for the PTT X.25 networks and their PADs was not harmonized with Y/11 Y/12 functional standards and, indeed, the work of T/31x. You will recall that it was my understanding that CEPT was harmonizing their networks and that this work was not part of the CEN/CENELEC/CEPT work. I took the view that it would be wasteful to continue the Working Groups activities without assurances that it was harmonized with CEPT work.
At that time I put forward the alternatives of:-
I have now received advice from some members of the Working Group which suggests the following action:-
Since I have received no alternative advice I shall be following the above proposals.
The draft ENV is now in a fairly advanced form and as long as the CEPT work does not cause a radical rethink I would not expect to see much change. However there are a fair number of points of detail that need resolution and I feel that these are quite suitable for resolution by correspondence so I hope that the situation will not delay the final document by much.
Yours sincerely
Paul Bryant.
You will recall the letter sent to BT to attempt to reduce the swinging volume tariff BT proposed to impose on EARN lines and the negative response received from BT.
In my latest discussions with Chris Hobson of the international regulatory department I have been led to think that BT has severe clerical problems in raising an invoice in the foreseeable future (if ever).
I am therefore proposing to continue to use the EARN connection without restriction until such time as I receive a bill I cannot pay.
The situation appears to allow BT to maintain their position of insisting on a tariff which will not weaken their position with respect to other organization while allowing us to continue.
For our part we must still maintain that we expect a bill at some time and will seek appropriate funding when it arrives. I also feel that we should soft peddle our criticism of BT on this issue and direct our energies towards CEPT to negotiate suitable tariffs for academic networks on a European basis.
I believe the situation to be as satisfactory as one could expect and I trust you have no objections to Rutherford taking the very small risk that BT will raise a bill.
Dear Mr. De Heer and Mr. Schouten,
Since the last meeting of the Working Group I have been told that CEPT is harmonizing the PTT X.25 networks and PADs across Europe. You should have received a letter from me outlining the problems that this presents to the Working Group. From the comments received from members the following action seems appropriate:-
I hope you will agree that in the circumstances this is the best course of action as I do not wish to waste peoples valuable time by coming to a meeting whose outcome may be overtaken by events within CEPT.
In practice I hope that the refinement of the current draft can continue and I will be circulating a new draft very soon which I hope you will comment on.
May I ask you to undertake the second proposal which is to tell us what CEPT is doing and to provide us with relevant documents? I would also value your comments on how you see the two activities interacting.
Yours sincerely
Paul Bryant.