Paul Bryant's Networking Correspondence
Dear Jean-Claude,
Here is my claim for my travel with respect to the EARN programme committee of February 25/26 in Dublin.
Air fare 117.00 UKL Hotel 29.55 IRL @ 1.1 = 32.50 UKL Total 149.50 UKL
Please make cheque payable to 'Rutherford Appleton Laboratory'.
Thank you.
With best wishes,
Paul Bryant.
This is version 2 which has been amended in the light of comment. Can we agree this version please. Alain - can you count the votes please. Also can we have views as to whether it is appropriate to issue this as an EARN Directive.
As requested by the EARN Executive I have redrafted the proposal considered on the 8 March Executive meeting. The Executive is invited to approve the proposal which will become our policy.
Proposal for an EARN Directive applicable to international nodes.
In order to operate an efficient network it is essential to manage it effectively. This requires that all changes in the technology used in the EARN backbone nodes (other than minor changes due to new releases of existing software or replacement hardware) must be proposed, discussed, and agreed by the Executive or Board of Directors before implementation. Such discussions should consider the technical detail, the effect on the complete user community, the management implication, the effect on other possible developments, and political considerations.
The EARN Executive will define and maintain a register of the relevant software and hardware.
X.25 network starting 1977 Now- About 50 DTEs mostly at 9.6 About 20 PADs about 250 terminals 7 PRIMES About 8 VAX A lot of other bits. Problem- bit slow - a lot of wire- complex
1983 installed a Cambridge Ring 4 Km the longest in the world A socket in every office Built on the 'petal principle' Worked very well
Equipment connected:-
6 Diamond word processors
5 GEC (home built 64K interface)
1 Camtec PAD
1 Camtec gateway
1 IBM non working interface
BUT
Expected MACE interfaces from Acorn
We still wait
Moral- NEVER BUT NEVER base plans
on equipment not demonstrable
Complete waste of money and resources
Difficult to get support for further
progress
Undertook study: Install backbone ethernet Expect local ethernets in 'Villages' Put Bridges between backbone and Villages
Built on a fibre optic 'star' BICC fibre optic repeater in CD DEC bridges run by CD EXCELAN monitor based on PC NO OTHER EQUIPMENT CD collects statistics, manages, and pulls the plugs
Village- a collection of closely related machines - VAX cluster or UNIX set PADs and terminals under control of villages Most traffic will be intra-village thus reduce backbone traffic Bridges will insulate villages and backbones
Want to use 'Pink Book' - Coloured Book over connection oriented (X.25) Allow DECNET as it is popular Allow TCP/IP as it is popular Phase out DECNET and TCP/IP as soon as possible
DEC - no problem - all protocols available GEC PRIME being phased out Spider PAD almost available Spider LAN/WAN gateway almost available UNIX systems have good TCP/IP
Use AUSCOM in house development No IBM product in View Provides coloured books and 3270 via PADs
All backbone and village LANs in place Many VAXes and IBM connected Lot of UNIX systems connected Traffic light
PADs from Spider are not yet satisfactory Spider gateway not available BUT can gateway via VAX Want IBM supported interface
Expect X.25 and PACX are at maximum size will start to reduce in 1989 Expect villages to look after own comms Expect less cable- less complexity Expect better performance Expect FDDI 100 M in 5 years as backbone.
A set of 'directives' and 'recommendations' have been developed which the EARN executive will issue after consultation and agreement with Board of Director members and Country Co-ordinators. They concern the way nodes operate their communications software. Comments are invited from the BOD and CC and should be made by 1st April 1988.
There have been difficulties with the transfer of large files and with congestions that occur from time to time in EARN. The Technical Group commissioned a proposal for alleviating these difficulties from Eric Thomas. His report was amended and agreed at the Technical Group meeting in Lisbon. The amended report was presented to the EARN Executive who have asked for comment to be invited from BOD and CC members with a view to the report's adoption. The 'directives' and 'recommendations' are reproduced below. The full document, which is large and contains explanatory information, will be supplied on request. Many members will already have a copy. BOD members should ask PEB@IB.RL.AC.UK and CC members should ask AUROUX@FRMOP11 for copies if required.
At a recent Executive meeting it was agreed that after suitable consultation 'directives' and 'recommendations' will be issued. These would apply to one or more of three classes of EARN nodes. These are 'international nodes', 'transit nodes' (nodes with two or more connections but excluding international ones), and 'end nodes'. It will be mandatory for EARN directives to be implemented on relevant nodes while recommendations are advisory. Members should bear this in mind when considering the proposals.
Comments of general interest may go the the BOD or CC distribution lists whilst other comments should be made to PEB@IB.RL.AC.UK or AUROUX@FRMOP11.
A strategy document was presented to the EARN BOD in Nice in April 1987 and largely accepted. This was also presented to the RARE Valencia meeting.
A subsequent meeting of EARN in Perugia considered a more detailed proposal. The main proposals were to install an X.25 backbone and to operate NJE/SNA over it on an interim basis.
DEC offered support in November 1987. This has led to a series of three meetings supported by DEC which have reconsidered the issues. The first meeting was in Geneva 18/19 February. The second was in Amsterdam 10/11 March.
The X.25 infrastructure ideas have not changed. That is, it is intended to set up a small number of switches with all the countries connecting to them. Northern Telecom is proposing to donate the equipment for this. Details are not yet available.
The proposals on the higher levels have now changed. It is proposed to develop NJE/OSI session/X.25. This product will be developed by Joiners associates on a VAX computer. It appears that this is about a months work. The possibility of a similar product on IBM machines based on GTMOSI is being looked at by the Italians.
A gateway is proposed between RFC822 and X.400 based on PRMD and the DEC x.400 product. This may also be undertaken by Joiners.
These DEC machines are now called 'G boxes'. It is proposed that each country will have a G box.
There are still several problems (as seen by myself).
There is a view that the only connection between the X.25 infrastructure and the current EARN will be via the G box. This is likely to be unacceptable to those who advocate SNA/X.25 as it will place two extra machines between countries and there is likely to be a loss of visibility through the G boxes.
It would therefore seem vital that the IBM machines also have the same protocol (NJE/OSI session/X.25) or the G boxes will be rendered obsolete by the expected popularity of NJE/SNA/X.25. The Germans already have an extensive network of this sort and the French and Italians appear to be following.
The proposed NJE/OSI session/X.25 preserves the idea of permanent connections between NJEs. Thus a G box will communicate with a small number of other such machines. This is because NJE does not understand the concert of a file or a call at the driver level. Thus staging will still take place. It would be possible to get over this by time out mechanisms and thus allow an end to end call between any suitable machines. This is attractive but fraught with problems. For example, what happens with messages? It may be unacceptable to set up a session for every message.
Almost no thought has yet been given to the addressing issues. In particular NSAPs.
It is expected that DEC will provide a G box for each country. DEC is also likely to provide a measure of support but this is undefined as yet.
The support of line costs is not yet known.
I met Kees Neggers in Amsterdam.
As a result of our previous meeting and of the Dublin meeting on EUNET, Kees is very keen to get the UK NL 64K line installed. He is willing to finance the complete line himself.
It is also clear that DEC would like Holland and UK to be trial sites for the G boxes.
Technically the line at this end would go into a switch which would allow the line to be used for a connection to the IBM, to the G box, to UKC, to JANET, and possibly to Ireland. The exact set of connections has yet to be agreed.
I propose that we should allow this line to be installed. This will take 4 months. In the mean time the details of the use of the line should be worked out.
I would expect that the manpower for this work will come from the expected recruitment.
Assume the 4 switch model will be adopted.
DPN-10 and DPN-15 Access Concentrators are not suitable for the 4 switches as they do not have sufficient capacity (only 200 or 400 dpps), have only 1 network link. They would be suitable for installation in countries with no X.25 infrastructure such as Greece or Portugal.
DPN-20 Network Switch is suitable for the switches as the capacity of 1000 dpps is sufficient (EARN minimum 800 dpps), has 6 network links (EARN minimum 2 and expansion to 4 desirable), has sufficient user access at 208 (EARN minimum 16). There are sufficient virtual circuits. The network line speed of 64K is initially sufficient but does not meet the aspiration of 2M line speeds although these are some way off.
DPN equipment offer, in addition to X.25, SDLC/SNA, 3270 DSP, 3270 DSHP, API, ANSI X3.28 and ISO poll. It is envisaged that EARN would want to offer these.
DPN equipment offers X.3/X.28/X.29. This service may be useful in a limited capacity on the backbone. It is envisaged that any such services would be provided in the attached networks.
DPN equipment follow the model of each sub-network being connected to other pier networks. The EARN model is to provide a hierarchical network with the top level being the international infrastructure and the second level being national networks.
From the documents it is unclear whether the addressing mechanisms of the DPN equipment can cope with this model. This would require the DPN equipment to switch traffic on the basis of the DNIC. It is unclear whether a set of addresses can be directed down a single access line or whether this can only be done via a network line and X.75.
EARN wants to use 14 digit addressing based on X.121. There will be no distinction between national and international addresses. DPN equipment (pg 178) appears to differentiate between national and international numbers by use of the 15 digit. It is unclear whether this digit really exists or whether it is only for X.28 use.
Routing is on the basis of the "area code" and "office code" parts of the DTE number. EARN wants to switch in the international infrastructure on the DNIC. This may not be possible.
DPN-30 50 are too large.
Howard Gibbs Northern Telecom
Paul H Dinsmore Northern Telecom
Bernard Harris DEC
O Jourgenson DEC
? DEC
Joe Chester EARN
Paul Bryant EARN
The visit was to undertake a technical evaluation of Northern Telecom network products for use in the EARN international infrastructure.
Northern Telecom are proposing the use of the DPN-100 'large capacity network data switch'.
Contrary to the name, the switch can be configured from very small systems up to very large ones. It supersedes their current equipment, the DPN-10, 15, 20, 30, 50 and earlier.
Northern Telecom have considered the EARN requirements and have independently arrived at a topology which is very similar to the topologies developed from Perugia. That is a network based on minimum number of switches and located to take advantage of low tariffs and taking account of known high traffic areas. Four switches at Montpellier, Switzerland, Holland and Stockholm are recommended.
The switch recommended is the DPN-100/20S.
The comparison with the EARN requirement is:
Perugia proposed that the running effort was about one man month a year assuming that the infrastructure was stable. This was thought to be too small but Northern Telecom suggest that very little operational effort is required. However, the analysis of statistics could absorb as much effort as one liked. Thus, a man year is a generous estimate once the infrastructure is in place.
The DPN products are just coming into productions and shipments should star in June.
The Northern Telcom DPN-100/20S equipment should meet the EARN requirements.
The X.25 network may be provided by the PTTs, other suppliers, or the community.
The criteria for the selection of a supplier is that the network must be provided with a non-volume tariff, at a price the community can afford, and to a acceptable technical specification.
There are a number of switch manufacturers capable of providing suitable equipment. Northern Telecom have offered to provide EARN with free equipment and support. This equipment meets the EARN specification (see below). Thus if EARN is to provide its own switches the Northern Telecom offer should be accepted.
(Speculative) IBM have offered to provide XI software and some upgrading of equipment to form an X.25 infrastructure. Unfortunately this equipment does not meet the throughput requirements and would also entail the use of current equipment. It is considered that the X.25 infrastructure should be free standing and not be dependent on any equipment providing other services.
(Speculative) The PTTs are expected to offer managed data networks. It has been established that some PTTs would be prepared to offer such services but they would be expensive compared with a community provided network. Such providers would be undertaking the installations tasks required. They would also provide lines and management. In the EARN case, the installation would be provided free by Northern Telecom. Northern Telecom are of the opinion that the fairly static and small network proposed would take a fraction of a man to manage. In addition the provision of leased lines is a task EARN is well briefed to undertake. The conclusion is that a managed network would be a comparatively expensive option.
(speculative) It would be possible for a PTT to manage the network installed by Northern Telecom. This would be likely to be an expensive option considering the static nature of the network and the low manpower required.
Assuming a community provided network a topology has been agreed which has been developed from the Perugia document. An independent study by Northern Telecom arrived at an almost identical topology.
The line topology is shown in figure 1.
Denmark------------+
Belgium-n----+ Finland-----------+|
Ireland-n---+| Iceland----------+||
UK------n--+|| Norway----------+|||
||| ||||
Holland---------64K n------------Sweden
| |
| |
| |
| |
64K n 64K n
| |
| |
| |
| |
| |
France----------64K n------------CERN
||||||||| ||
Cote d'Ivore+|||||||| Austria-----------+|
Greece-------+||||||| Germany--64K-------+
Italy --------+||||||
Israel---------+|||||
Portugal----n---+||||
Spain------------+|||
Tel Aviv----------+||
Turkey-------------+|
|
|
USA
Fig.2 X.25 infrastructure
n=new lines
All lines at 9.6K unless otherwise stated.
Switches will be located at Nijmegen, CERN, Montpellier, and Stockholm. This is subject to negotiations with the sites.
Where possible existing lines will be used at the existing speeds. The connections between Nijmegen, CERN, Montpellier, and Copenhagen will be at 64K as soon as lines are available from the PTTs.
The initial lines will include the lines between the switch sites. As these are new lines with no existing services they will be dedicated to inter switch traffic from installation.
The remainder of the connections will be required as and when countries decide to join the infrastructure. A schedule for connections will be drawn up in consultation with countries.
The requirements for an X.25 switch are:
The requirements for X.25 DCE connections are:
The LCN allocation will be:
Permanent virtual circuits- groups 0 through 3 One way incoming- groups 4 through 7 Two way- groups 8 through 11 One way outgoing- groups 12 through 16
The use of 'one way incoming' and 'one way outgoing' groups are for further study. The range of LCNs for 'two way' will be contiguous from the lowest. The range size will be a subscription time option. The DCE will allocate LCNs is ascending order. The DTE is recommended to allocate LCNs in descending order. PVC LCNs between an IBM SNA node and a switch will be the same as the SNA network number of the local SNA network. (See 'Use of SNA on the EARN X.25 infrastructure). In brief, the network number will be obtained by listing all the EARN countries in alphabetic order of their two character ISO country code and then numbering then from 1 upwards). The PVC number between switches will be the 'lowest SNA network number'*32+'highest SNA network number'. The allocation of PVC logical channel numbers for other purposes, for example DECNET, is for further study.
The CCITT X.121 Recommendation will be used.
DTE number will be:-
+------------+----------------------------------------------------+ | DNIC | NTN (Network Terminal Number) | | 4 +-----------------+---------------+------------------+ | digits |4 digit site code|4 digit machine|2 digit subaddress| +------------+-----------------+---------------+------------------+
The DNIC shall be the three digit country code as specified by CCITT followed by a digit to be selected by the country but 0 is recommended.
It is recommended that the '4 digit site code' should be selected by the country and that it is based on telephone area codes.
It is recommended that the '4 digit machine' should be selected by the site but a country may wish to impose further recommendations.
The optional '2 digit subaddress' will not be policed by the network and the use to which these digits are put is not defined.
The DTE numbers shall be registered with a network management centre.
For interim periods countries may have more than one connection to the X.25 infrastructure to enable any transitions to be safely accomplished.
Permanent multiple connections from a country to a switch will require finance for relevant additional equipment from the country.
1/4 man year is required at each switch site which will ensure the switches are operational and will manage any required reconfiguration of hardware.
One man year is required at the management centre which will monitor performance, provide statistics, re-configure software, allocate DTE numbers.
Some manpower will be required at sites which do not have SNA or X.25 expertise.
Capital Recurrent
UKL UKL
Four switches at 10000 UKL each 40000 4000
Line to complete square (assuming
current proposed changes take place) 20000
1/4 man year switch support per site 30000
One man year network management centre 30000
Costs for connection of computers and other equipment to the infrastructure will be met from national resources. This will include any international line changes.
The finance required will depend on the supplier. An estimate is that the cost of connections should be the leased line costs plus 4000UKL. This estimate is based on the cost of private switches doubled to take account of their management.
The actions required to implement the the X.25 infrastructure are:-
The actions required for the provision of the X.25 infrastructure are:
Country ISO 3166 2 and ISO 3166 X.121
3 character code number DNIC
Austria AT AUT 040 232
Republic of Austria
Belgium BE BEL 056 206
Kingdom of Belgium
Denmark DK DNK 208 238
Kingdom of Denmark
Finland FI FIN 246 244
Republic of Finland
France FR FRA 250 208
French Republic
Germany DE DEU 280 262
Federal Republic of Germany
Greece GR GRC 300 202
Hellenic Republic
Ireland IE IRL 372 272
-
Israel IL ISR 376 425
State of Israel
Italy IT ITA 380 222
Italian Republic
Ivory Coast CI CIV 384 612
Republic of the Ivory Coast
Luxemburg LU LUX 442 270
Grand Duchy of Luxembourg
Netherlands NL NLD 528 204
Kingdom of Netherlands
Norway NO NOR 578 242
Kingdom of Norway
Portugal PT PRT 620 268
Portuguese Republic
Spain ES ESP 724 214
Spanish State
Sweden SE SWE 752 240
Kingdom of Sweden
Switzerland CH CHE 756 228
Swiss Confederation
Turkey TR TUR 792 286
Republic of Turkey
United Kingdom GB GBR 826 234
The United Kingdom of Great Britain and Northern Ireland
Dear Jean-Claude,
Here is my claim for my travel with respect to the EARN programme committee of March 22 in Paris.
Air fare 170.00 UKL RER 25-80FF @10.5 2.46 UKL Total 172.46 UKL
Please make cheque payable to 'Rutherford Appleton Laboratory'.
Thank you.
With best wishes,
Paul Bryant.
I have updated the domains paper according to the comments. I did not understand Harri's point that 'there must be two Internet nameservers serving the domain and all national gateways.....'. Am I right in thinking that one nameserver is DOMAIN NAMES and the other is some nameserver within the national network? Unfortunately, I cannot see that we can avoid the % hack and source routing. If you take the UK, ARPA is registered as going via the UCL gateway and the only way to go via EARN is the % hack. We also cannot register UK in SRI-NIC as it is already there as routed via UCL. Or have I missed something.
RFC 822 mail systems have been steadily adopting domain addressing.
Details of domain addressing can be found in RFC 920.
This proposal is that EARN should encourage the use of domain addresses. The benefits are:
For historic reasons the use of domain names in the USA has tended to be based on networks and has resulted in top level names of the form ARPA, COM and so on.
In Europe the desire is to adopt the two character ISO country codes as the top levels such as FR and CH.
In Europe the objective would be for the address of a mail box to be the same regardless of which network it is on and the access route. In fact a mail box would have an RFC 822 form and an X.400 form. RFC 987 provides the mapping between the two.
EARN would recognise within Europe 20 or so top level domains, one for each country. The other components of an address would be a matter for resolution within a country and agreed with all the interested parties. In some cases decisions have already been taken.
The proposal requires that a single system exists in each country which will accept mail addressed to the country. It may be that initially such mail will be rejected until systems within the national EARN can accept the addresses and/or relevant relays to national networks have been established.
The top domain names will be in the NETSERV file DOMAIN NAMES as .FR .CH and so on. The top level country codes should be registered in SRI-NIC. This may not be possible in some cases as the code will have already been registered by some other organisation.
Countries shall set up a single mail service capable of receiving mail addressed to the country domain.
Countries shall if possible register their country domain name in SRI-NIC.
Countries are encouraged to extend the use of domain addresses within their country and to provide relays to appropriate national mail services.
Most of the comments are linguistic or pedantic and can be ignored. Important points are indicated by ***. It looks to be shaping up well. My main worry is that several statement are not very well supported and may be attacked. Of particular concern is that the technical details of NJE/OSI are almost non existent. This means that the reader has difficulty deciding whether it is a good idea or not. To be more specific, is it intended that a call will be set up per file, is it intended that every node will send files directly to other nodes? If these are not possible then NJE/SNA has some significant advantages which the SNA advocates will undoubtedly grab hold of. The SNA advocates may well argue that it would be simple to mount NJE/SNA on a VAX. Our text does not show this to be difficult or impossible (perhaps it is easy). The sizing of the G-box worries me - I would like to see some calculations to show that a G-box will have sufficient CPU cycles and sufficient disc space, some of this work did come out at the meetings but is not recorded. WE MUST REMEMBER THAT OUR READERS WILL NOT HAVE THE PRIVILEGE OF HAVING ATTENDED THE MEETINGS AND MAY BE HOSTILE. I have not identified these as I have difficulty in producing appropriate text. Please excuse me playing the Devils advocate. Paul.
My X.25 section follows. Please feel free to "do a job" on it. I am too close to it to see its faults.
The section "The brief of the team" and "Objectives for the EARN.." contain very similar material. The "The brief.." section could be omitted with no loss of information.
Section "The brief.." - penultimate line - delete first "to".
Section "Objectives for.." - you switch to using "group" instead of "team". I think this occurs else where and a "global" is needed.
Section "Objectives for.." - line 6 - sentence starting "In addition.." - the sentence is too long - delete "In addition", end sentence after "necessary" and replace "and to" by "They were asked". Suggest replace "pertaining" by "relevant" - I am thinking of the non English reader.
Section "Objectives for.." - "2. Define.." - The first sentence is very long and difficult to understand. The phrase "in particular.." lacks a verb. The two phrases seem to be saying the same thing twice. (It looks like a committee invented sentence). Suggest "." after "EARN" and replace second phrase by sentence "In particular, the path should lead to X.400 becoming the primary network mail service.". The omitted text seems to add nothing.
Section "Objectives for.." - "3.Provide.." - line 2 - replace "and" by ",". A list should be separated by "," except (optionally) the last.
PART 1- line 4 - use of "today" and "existing" is a tautology - suggest delete today.
PART 1 - "1. Mail Services.." - In previous numbered sections you do not have headings, in this one you do. Also, suggest deletion of "at the moment" it is a tautology as three lines above "Existing" is used. In line four, "also" should be deleted as it implies that BSMTP has been used for something else. In the fifth line "code" seems inappropriate. I do not understand the use of "for example" at the end of the paragraph". It implies that there are yet more mail protocols getting "considerable use" which is incorrect. Suggest "for example" is deleted.
PART 1 - "2. Unsolicited.." -Suggest deletion of "at present" - tautology again. Suggest "," after "formats" as the "and"s are ambiguous - even better delete all "formats" and "and"s except last and add "formats".
PART 1 - "Specific applications" - Suggest replace "these generic" by "the above". I do not think "generic" is the appropriate word to use.
PART 1 - "Migrating these.."- line 3. Suggest "," after "The conclusion was that" and in line 4, Suggest replace "that" by "," - the sentence structure is not easy to follow. Line 5. suggest "In addition to this" is deleted - it is pure padding. In line 7 I would delete "probable" as it seems fairly "certain" to me.
PART 1 - "The introduction.." - line 3. You change person. Suggest "..bring an OSI network any closer.".
PART 1 - "In Perugia..". Line 2. Tense. Suggest replace "we need" by "needed". The second sentence causes confusion as the reader assumes that NJE/OSI is rejected. Suggest sentence is started with "But" and "solutions" is replaced by "NJE over OSI" and replace "NJE over OSI" by "it". The sentence is long, suggest a stop after "place" and delete "and". Last line, "as a solution" is difficult to understand as it is unclear what it refers to. Suggest it is deleted.
**** PART 1 - "In Perugia.." - last line. Do you really mean it has to implementable on "all" hosts? How do you intend to check this? and if the check fails does the strategy fail? Suggest "all" is replaced by "all major".
PART 1 - "NJE as an..". Line 4. Suggest replace "platform" by "systems". May confuse the non English reader. (Never use an obscure word when a well known one will do).
PART 1 - "NJE over OSI.." Second paragraph - is it "Jnet" or "JNET"?
*** PART 1 - "NJE over OSI..". "The Jnet.." - Line 2. We are asking for trouble by including SNA here. If in fact Jnet runs over SNA then the SNA solution works with no effort at all! It should be pointed out that this is not NJE/SNA-session/X.25. Steve assured me that it was not.
PART 1 - "The Jnet". Line 3. Suggest you use the more usual term "Jnet line driver" unless this is a new technical term whence use "". In line 4 I worry over the word "standard" (I always do), how about "supported". In the last line, to be consistent "X.25 interface" should be in "". I also think that "interface" should be "service".
PART 1 - "Joiner have..". Replace "Joiner" by "Joiner's". There are too many "solutions" in the sentence. Suggest replace "and this solution" by "which". Also the use of "and"s makes the sentence ambiguous, suggest replace final "and" by ".".
PART 1 - "NJE over OSI Session on IBM hosts" - line 3. If you define GTMOSI you should also define OSNS and OTSS. Perhaps these definitions should be in a glossary. In last sentence "actually" can be omitted with no loss.
*** PART 1 - "(a).." - In line 3 I have problems with SNA. SNA and OSI are structures with many parts and the parts referred to should be stated here. My knowledge of these bits and pieces is not enough to suggest better text.
PART 1 -_ "This proposal.." - line 2. Should this be RSCS V2?. Line 3, suggest "RSCS or JES2" is replaced by "them". In line 4, "someone" seems a bit loose, suggest delete "someone in" and add "personnel" after "IBM".
PART 1 - "(b)..". Suggest delete "In this case" and "This time". Tautology.
PART 1 - "Conclusion" - line 2. Add "the" before "development". In line 8 I do not understand why this strategy "ensures interoperability of OSI applications". I think what you are trying to say is that this strategy, in exercising the OSI protocol products, will ensure that these products interoperate thus lessening any interruption problems with new OSI applications. Last line - typo - "international".
PART 1 - "Recommendations". This section can be read as implying that NJE is an ISO application. Suggest "as an OSI application" is deleted. Line 4, should "Continuity of Service" be in lower case and perhaps be in "".
PART 2 - "General Commitment..". Replace "at this point" by "currently".
PART 2 - "There is great..". This paragraph is ambiguous. Does the second sentence refer to directory services only or a wider range of services. It is also unclear as to whether you are referring to directory services to "drive the network" or directory services to look up peoples names.
PART 2 - i) ii) and iii). Use 1. 2. and 3. as in other places. In i) tense, Suggest "If all the existing services can be provided using OSI applications". Similarly for ii) and iii).
PART 2 - "Gateway Services" - line 3. Suggest replace "of some" by "between some". Line 3, "," after "example". Line 5, either delete "gateway" or make plural. Line 5, replace "This" by "The".
PART 2 - 2.1. You use a) i) 1. and 1) in a fairly arbitrary way. Also, i) a few lines above has no title and in this section it has. Can we be consistent.
PART 2 - "X.400 Pilot Project". There are two lists in this section, the first has "-" before each item, the second has not.
PART 3 - "The principle .." - line 2. Suggest "." after "initially" and delete "and". Suggest "," after "project".
PART 3 - "Recommendations" - line 6. Replace "need" by "needs". Line 8, suggest replace "conformance....." by "conformance and interoperability testing of the implementations". This reads better and also you cannot test a solution until implemented.
*** PART 4 - line 9. I do not understand "X.400 address resolution data". How about "Appropriate X.400 address resolution procedures"? Why is an X.400 relay required?
PART 4 - "3. Communications..". I would not mention X.29 in the same class as BSC etc.
PART 4 - "In addition.." - "," after In addition. Or even better delete it.
PART 4 - very last line. Is it the "Board" or "Executive" which decides?
Joe Chester EARN administration (Dublin) Paul Bryant EARN technical director (Rutherford, UK) Michael Hebgen (Heidelberg University, Germany) Jose-Maria Blasco (GMD Bonn, Germany) Pann Pictikainen (Helsinki University of Technology, Finland) Harri Salminen (FUNET, Finland) Jukka Korpela (Helsinki University of Technology, Finland) Sitki Aytac (Ege University, Turkey) Olivier Martin (CERN, Switzerland) Jean-Luc Delhaye (CNUSC, France) Dominique Pinse (IBM Paris, France) Dominique Dumas (CNUSC, France) Eric Thomas (CERN, Switzerland) Berthold Pasch (IBM-ENC Heidelberg, Germany) Peter Sylvester (GMD-Bonn, Germany) Manfred Bogen (GMD-Bonn, Germany) Ulrich Giese (KUN-Nijmegen, Holland) Udo Meyer (GSI Darmstadt, Germany) Wolfgang Hake (University of Bielfdal, Germany) ? Jos Wennmacher (KUN-Nijmegen, Holland) Kieran Carrick (University College Dublin, Ireland) Miguel Campos (Universidad Barcelona, Spain) Peter Streibelt (IBM HQ, Germany) Wilfreied Racke (IBM ENC Heidelberg, Germany) Rene Florizoone (University Leuven, Belgium) Wim Brems (University Leuven, Belgium) Guenther Schmittner (University of Linz, Austria) Gerard Pitteloud (University of Zurich, Switzerland) Moshe Barak (Technion, Israel) Alain Auroux EARN manager (France) Daniele Bovio (CNR-SIAM Milano, Italia)
Dear Paul,
Many thanks for allowing me to attend the X.25 switch course which was very useful.
In discussing the EARN requirements with some people on the course I think we may have a problem with the scheme we propose using. Unfortunately the experts were not familiar enough with the switch to give me a definitive answer. I attach a document which examines how we envisage using the Northern Telecom equipment which draws attention to the problems. I am wondering if at some time I could speak to one of your experts to sort out the problems.
I shall be away until 1 June when I will give you a ring to decide what to do next.
I have just seen the press release given in Sweden which outlines the offer from Northern Telecom. I am confident that we can now go ahead with confidence to meet our users requirements.
Yours sincerely
Paul Bryant. (EARN technical director)
There will be an international X.25 network with one or two connections to national networks, gateways, or special facilities within a country. This will be based on Northern Telecom switches.
X.121 like addressing will be used with each country having a four digit DNIC. The international network will have its own DNIC (2000) but this will only support the network management machine and possibly a few special services such as directory services.
Each country will be connected to the international network by methods depending on local circumstances. These have yet to be determined with each country. The methods of connection are:
The national network will have Northern Telecom switches. The international and national networks will be a single X.25 network exhibiting the DNIC of EARN and the national DNIC. The complete network will be managed by EARN from a single management centre.
There appears to be no problems with this scheme apart from the country having to comply with the DTE address length used in EARN.
It is likely that the less developed countries may opt for this scheme.
The national network will have to exhibit X.75. The country will have full autonomy over their network apart from having to accept calls containing the country DNIC and generate address for other countries with their DNICS.
A country may have to employ some means of accommodating the EARN addresses.
As far as is known no country has a private network which exhibits X.75.
Many national networks cannot accept X.75 connections but can cope with a number of DTE addresses via a single connection.
It appears that the Northern Telecom switch is unable to provide this type of connection.
A country may have to employ some means of accommodating the EARN addresses.
The UK network JANET is in this category. It is unable to accommodate X.75 and has a DNIC of 0000. It appears that the Northern Telecom switch can deal with the DNIC problem by call redirection. Remember that the UK will be known in the rest of EARN by the X.121 UK DNIC.
Some G-boxes may be connected to national X.25 networks (methods 2.1/2/3). Such connections are outside the scope of this paper.
G-boxes may be connected directly to the international network. This is no problem if that is the only connection to the country. However there may also be a connection to a national network which would be via an X.75 connection. Unfortunately the Northern Telecom switch cannot select a few addresses with a given DNIC for X.25 connections while directing the rest with the same DNIC to an X.75 port.
A solution is for a country to have two DNICs differing only in the fourth digit. The first would be for direct X.25 connections and the second for an X.75 connection.
Many countries will initially require a G-box directly connected to the international network.
A number of countries have networks which are not X.25. In these cases relays and/or gateways will be provided by the countries. The problems have not been studied but are similar to the connection of G-boxes and the equipment may, in fact, be modified G-boxes.
Finland is in this category as they are operating a national TCP/IP network.
Several countries are, or expect to use, the public networks. It seems unlikely that the public services will offer X.75 to customers. With X.25 connections some administrations will switch groups of numbers but are unlikely to direct the whole range of EARN DTE numbers to a connection. In any case EARNs address scheme is not orthogonal to the public scheme.
The only scheme appears to be to use a network level gateway. This has yet to be studied. It would appear that such a gateway would only require an X.25 connection into the international network which would be no different to a G-box connection.
Some countries may wish to use network level gateways irrespective of whether they use the public networks. So far no implementations of such gateways are known. Comments in 2.6 apply.
The EARN address scheme based on X.121 has been well received. It has the advantage that:
If the Northern Telecom switch cannot switch a group of connections via an X.25 rather than X.75 line then there are a number of options:
Cyril.
I felt that I left just as the s**t was about to hit the fan. I thought that the meeting was a bit quiet and that the press release had not gone home. Did you get away with your skin?
I was hoping to have a discussion with you on the next stage and the recent developments.
There still seems an immense gulf between what I thought we should be trying to do and what the meeting is doing. I thought we were trying to see what room for co-operation there was between various network suppliers like ourselves. We still seem to be designing THE COSINE infrastructure. This is an indication that COSINE is failing to give any confidence that it will provide anything in the foreseeable future. Our terms of reference discussion seems to have achieved no redirection of effort.
At no time did we address the question of just what co-operation could take place. In fact EUNET said almost nothing. I got the feeling that they had lost interest as there seems nothing in it for them. Kees seems unwilling to discuss practical co-operation or perhaps we should press it more?
Questions like whether we will use the PPSDN, MDNS, or private is really a waste of time if EARN will have put in a private network before the WPA decides or recommends on such issues.
The papers from Nick and Ian contained nothing new and really were statements that I have seen in many other documents. I have to agree with them as they contain just about nothing of substance. They commit no one to anything.
How on earth is the COSINE Policy group going to make anything out of it. We have formed absolutely no collective (or other) opinions or recommendations for our organisations. The reason is obvious in that the moment we say what we really think the meeting will dissolve into religion. Our religion that we want to put in this network like yesterday, their religion that it must be run by the PTTs (or colleagues) and be THE COSINE infrastructure under COSINE (or colleagues) management.
I am inclined not to believe that Nick has not got 1 Mecu. It is rather like the Alcatel offer. In any case 1 Mecu is just about enough for some further study.
I admire your attempt to list some future work items. I am doubtful whether WGA can address those items rather than EARN itself.
I ask just how much effort can we afford to put into this charade. They think they have stopped us and control us (or at least hope that is the case). We think we are poised to continue (or at least hope that is the case). WPA, to my mind, is an attempt on COSINE's side to control us and on EARNs side to keep friends with then and try and put across our point of view. I am now inclined to put minimum effort into WPA.
But, we need to have our co-operation with EUNET et al. We will not get this through WPA. I believe we should now have informal contact with other networks to further what should have been WPA's aims.
It is getting important that we start to plan the next stage. To some extent depends how negotiations are getting on with DEC. In principle we should get the programme committee into action.
A hoped for result of putting in a local area network was to reduce costs and provide a better service. The reduced costs were expected to come from the decline in the existing PACX and X.25 services.
The support of a single network should in principle require less man power. Should allow more resources to go into a network with a better potential.
We need to see if this true assuming that the LAN equipment does provide the services and is reliable. There are some basic questions to address:
As a first stage we need to collect information on which to decide on a philosophy.
Dear Jean-Claude,
Here is my claim for my travel with respect to the EARN/IBM and EARN programme committee meetings on 5/6/7 April.
Air fare 243-00 UKL
Taxi Frankfurt/Heidelberg see claim by Joe Chester
Train Heidelberg/Frankfurt 19-40 DM
Train Frankfurt/Airport 1-70 DM
subtotal 21-10 DM 1 UKL=3DM 7-03 UKL
Bus Amsterdam to hotel 11-00 NLG
Bus hotel airport 11-00 NLG
Hotel 180-00 NLG
Breakfast 8-50 NLG
subtotal 210-50 NLG 1 UKL=3.5NLG 60-14 UKL
Total 310-17 UKL
Please make cheque payable to 'Rutherford Appleton Laboratory'.
Thank you.
With best wishes,
Paul Bryant.
I reported on progress.
I have now visited Northern Telecom who are proposing to give EARN 4 DPN-100/20S switches.
The topology proposed by NT is remarkably similar to my Perugia proposal and is based on 4 switches with all countries connecting to a switch (no sun-switches). The switch sites are Holland, Sweden, CERN, and France located at the obvious sites. The UK now has no switch. I believe this to be sensible as Holland is a low tariff area with better connectivity to other countries. Initially, apart from Germany, the connections would be at 9.6 and the interswitch links are 64K
The switch seems to meet all the EARN requirements. In fact they are an overkill with the packet rate being a staggering 3500 date packets per second. The DPN100 is a modular range of switches which can go from switches of a few lines up to many thousand. The reliability is 99.999%.
I have confidence that these are really excellent switches and, as they will be free and maintained for a long time by NT, it is an offer that cannot be ignored.
NT say the network will require a small fraction of a man to operate it. As suggested by myself at Perugia and thought unrealistic. Kees Neggers is very keen that the Dutch PTT should do this. I believe (in concert with Dennis) that we could not afford it. It is blatant Dutch empire building. I believe EARN should employ someone (preferable from outer space) to manage this network in a not nationalistic way. Fortunately NT will do all installations and setting up.
Kees is also very keen that we should wait for the various COSINE working groups to report but the feeling is that we should not let such considerations slow our development. There was also some feeling that the network should be a co-operative venture between EARN, EUNET, and COSINE. Here again we felt that this would virtually stop progress while the politics got sorted out and we would be in danger of EARN vanishing under a heap of other organisations.
IBM are offering X.25 based on the XI product in 37xx front ends. This has an equally staggering performance of 200 data packets per second which puts it out of court. In addition, we want the X.25 to be 'free standing' and not tied to any existing equipment which is doing other jobs. Pax vobiscum XI.
We will be approaching CEPT for a managed network and we will copy the letter to most PTTs. We expect that the PTTs will not be in a position to provide such a service at a price we can afford and cheaper than the leased lines plus very little.
The DEC promoted meeting (2 so far) are keen on the use of NJE/OSI session. This has attractions and we have now established that it is possible with little effort. In fact Joiners are doing it on the VAX and the IBM code is reported to be 50 lines of REX. It is attractive.
Joiners will also be doing an X.400 to RFC822 for us which is very attractive.
The SNA contingent are not lying down and I have persuaded them to produce an alternative proposal based on Perugia ideas. I have also stuck a spoke in their wheel by getting the Executive to agree that any use of 'new' methods of working MUST be approved by the Executive. My friends in Germany are about to go into orbit as they regard EARN as a set of bilateral agreements. The blue touch paper is lit on the 'who controls EARN' rocket.
It is now proposed to get IBM and others to propose how the SNA migration strategy would work and we can then select one route and/or the other. I am currently on the side of the NJE/OSI but it would be a tragedy if the network got split between the two methods.
The bones of the full proposal are now being put together. I am doing the X.25 parts and will probably go on a course with NT to learn more. The higher levels are being penned by Kieren Carrick of UCD.
I am now happier that progress is being made and that finance and support are in being.
The meeting will start at 9.00.
1 Election of a reporter
Following the resolution at the Lisbon meeting a reporter will be elected who will produce a report of the meeting which, after agreement with members, he will present to the Executive.
2 Minutes of the Lisbon meeting
These have been circulated via EARNTECH. The reporter (Peter Sylvester) attended a meeting of the Executive and presented the report and will report on the Executives response.
3 Matters arising from the minutes
4 Report on resolutions from Lisbon to the Executive
5 Transition to the use of ISO protocols
This item is for information and comment. There have been a number of meetings and working groups which have been developing a transition plan. This work is still progressing but has reached a stage where implementation can start. The Board is expected to take decisions at their meeting in Izmir. There are still many details to be decided. At this meeting the current thinking will be presented and comments invited.
Documentation on transition is still under production and will hopefully be available at the meeting.
Any comments from the meeting will be given to the Board at their meeting.
6 Change of responsibilities
As a result of the reduction in IBM support, the Lisbon meeting agreed to the reallocation of some tasks. The meeting will review the new responsibilities and recommend any changes required.
7 Statistics and topology
A discussion of progress since the Lisbon meeting.
8 Report on the "Operations Group" (Alain Auroux)
This group was set up under Alain Auroux in order to co-ordinate the operation of EARN.
9 Domain names (Paul Bryant)
A revised proposal will be circulated shortly.
10 Adoption of two character ISO country codes (Eric Thomas)
A clarification of the current state of progress.
11 Nodes file format (Berthol Pash)
There has been discussions in NODMGT-L and there is a willingness of BITNIC to make improvements to the nodes file. Since there will be a working group meeting before the next SHARE meeting at which there will hopefully be a resolution on future formats and node management in general - EARN should present its views and ensure that the new formats follow them. The particular issues are:
12 Contacts with BITNET (Berthol Pash)
There is a suggestion to send representatives of each network to the technical meetings of the others. So far, such visits have been unofficial and representatives are not explicitly authorised to speak for their organisations.
It is proposed that an official technical representative be selected and that the BOD should support his attendance at BITNET meetings. In particular expenses would be met.
One of his tasks would be to present the node management requirements of EARN at the next BITNET meeting.
13 Discussion on the operation of "Task Groups"
The Lisbon meeting decided to set up a series of Task Groups to resolve various problems. The meeting should review their effectiveness and recommend any changes.
14 Report on the NICE 2 (Peter Sylvester)
A status report on the NICE 2 directory project and a short overview of X.500 and how BITEARN NODES can be integrated into such a directory.
15 Organisation of "Birds Of a Feather" groups (Eric Thomas)
16 Date of next meetings
It is proposed that the next meeting will be in Finland.
A proposal for the subsequent meeting is requested.
16 Any other business
A proposal for the EARN BOD
Izmir, Turkey 21/22 April 1988
EARN would like to thank the many people
who have contributed to this report
1
WHAT ARE WE TRYING TO ACHIEVE?
A coherent
assembly of
networks across
Europe
Problems: Many network technologies
Many pressures
PTTs, CEPT, RARE, COSINE
IBM, Super computers, DEC
Finance
Manpower
2
LOWER LAYER STRATEGY
Use of ISO protocols
X.25 only viable option
ISDN - insufficiently developed
TP4 services - insufficiently developed
3
INITIAL STRATEGY
To interconnect existing networks
(or parts of EARN)
Scheme 1. bilateral connections
Scheme 2. common infrastructure
4
BILATERAL CONNECTIONS
(as used by EARN and PTTs)
National National
Network Network
1 2
National National
Network Network
3 4
Advantages- Use current topology
Little management
Disadvantages- Many interfaces
Difficult to plan
Subject to national whims
5
COMMON INFRASTRUCTURE
National National
Network Network
1 2
International
Infrastructure
National National
Network Network
3 4
Advantages- "Standard interface
Easy to plan
Needs central management
Disadvantages- Finance needed
Virtually a new network
6
PROPOSAL
Proposed that a centrally
managed infrastructure be
set up
7
OPTIONS FOR THE PROVISION OF THE INFRASTRUCTURE
* Use of PTT services
* Use of PTT managed service
* Use of EARN equipment managed
by contractor
* Use of EARN equipment managed
by EARN
8
USE OF PTT SERVICES
Unlikely to provide non-volume tariff
Modest quality of service
9
USE OF PTT MANAGED SERVICE
PTTs being approached
Unlikely that PTTs will be able to
provide a service in near future
No costs available
No performance figures available
10
USE OF EARN EQUIPMENT MANAGED
BY CONTRACTOR
Various contractors are being approached
Costs not known
Switches from Northern Telecom.
11
USE OF EARN EQUIPMENT MANAGED
BY EARN
Two suppliers (NT and IBM) identified
Topology has been examined
Management aspects have been examined.
12
TOPOLOGY
Small number of switched as-
switches expensive
better performance
Larger number of switches as-
reduce line costs
better resilience
Conclusion -
About 4 switches.
13
PROPOSED TOPOLOGY
Sweden (Nordics)---+
Belgium-n----+ |
Ireland-n---+| |
|| |
|| |
UK (RAL)--------64K n------------CWI (Amsterdam)
\ /
\ /
\ /
\ /
\ /
/ \
/ \
/ \
/ \
/ \
France----------64K n------------CERN
||||||| | ||
Ivory Coast-+|||||| | Austria-----------+|
Greece-------+||||| | Germany--64K-------+
Italy --------+|||| |
Israel---------+||| |
Portugal----n---+|| |
Spain------------+| |
Turkey------------+ |
|
x
|
USA
n = new lines
d = dial up
x = not X.25
All lines at 9.6K unless otherwise stated.
14
NETWORK SPECIFICATION
Infrastructure must be independent-
not dependent on service computers
Centrally managed-
infrastructure managed from a single centre
no local management
central collection of statistics etc.
15
SWITCH SPECIFICATION
Must provide X.25 (84) for ISO network service
Must provide SVCs and PVCs
Must provide up to 20 DCE/DTE interfaces
Must provide up to 500 SVCs
Must provide up to 200 PVCs
DCE interfaces up to 64K (2M later)
Must provide management/statistics etc.
Must support EARN address scheme X.121
Must connect to other X.25 networks
Must support 800 data packets per second
Must be maintainable in relevant EARN countries
Must support equipment conforming to Y11 and T31
16
NORTHERN TELECOM SWITCHES
Based on Northern Telecom DPN-100/20S
Meets all requirements
Also-
switches 3200 data packets per second
supports 500 calls per line pair
can manipulate address for connection to
other networks
extendible to 30,000 lines and 30,000 dpps
wide range of facilities- hunt group,
call redirection, closed user group,
X.25 dial out, NUI validation, 3270 bisynch,
SDLC, SNA, XXX, X.75
requires VAX/VM or IBM/VM for management
used by many PTTs
availability claim 99.999%
17
X.25 ADDRESS STRUCTURE
DTE number will be:-
+------------+----------------------------------------------------+
| DNIC | NTN (Network Terminal Number) |
| 4 +-----------------+---------------+------------------+
| digits |4 digit site code|4 digit machine|2 digit subaddress|
+------------+-----------------+---------------+------------------+
Mandatory DNIC - 3 digit country code + 1 digit country selected
Highly recommended NTN - 10 digit structure
Recommended - 4 digit site code - country selected
4 digit machine - site selected
2 digit subaddress - machine selected (use not
recommended)
DTE numbers are registered with "management centre".
20
CODES
Country ISO 3166 2 and ISO 3166 X.121
3 character code number DNIC
Austria AT AUT 040 232
Republic of Austria
Belgium BE BEL 056 206
Kingdom of Belgium
Denmark DK DNK 208 238
Kingdom of Denmark
Finland FI FIN 246 244
Republic of Finland
France FR FRA 250 208
French Republic
Germany DE DEU 280 262
Federal Republic of Germany
Greece GR GRC 300 202
Hellenic Republic
Ireland IE IRL 372 272
-
Israel IL ISR 376 425
State of Israel
Italy IT ITA 380 222
Italian Republic
Ivory Coast CI CIV 384 612
Republic of the Ivory Coast
Luxemburg LU LUX 442 270
Grand Duchy of Luxemburg
Netherlands NL NLD 528 204
Kingdom of Netherlands
Norway NO NOR 578 242
Kingdom of Norway
Portugal PT PRT 620 268
Portuguese Republic
Spain ES ESP 724 214
Spanish State
Sweden SE SWE 752 240
Kingdom of Sweden
Switzerland CH CHE 756 228
Swiss Confederation
Turkey TR TUR 792 286
Republic of Turkey
United Kingdom GB GBR 826 234
The United Kingdom of Great Britain and Northern Ireland
The EARN international network will have the identification-
EARN ER ERN 200
21
CONNECTIONS TO A COUNTRY
Eventually- one connection to a national
network (more than one for capacity)
Initially- a connection to an "NJE relay"
and "RFC822/X.400" relay
Interim- 2 connections by use of dual channel
modems, X.25 concentrator, or small switch.
21
CONNECTION TO NATIONAL NETWORKS
Option 1- Country buys Northern telecom switches
and shares management.
Advantage- Interconnection transparent to user.
Gives best possible international
quality of service.
Very easy to manage and organise.
Disadvantage- Country losses autonomy
22
CONNECTION TO NATIONAL NETWORKS
Option 2- Country has other than Northern Telecom
switches but shares address space
Advantage- Interconnection transparent to user.
Some autonomy for country
Disadvantage- Some loss of quality of service.
Some loss of country autonomy.
24
CONNECTION TO NATIONAL NETWORKS
Option 3- Country has a different X.25 address
space.
Advantage- Full autonomy for country.
Disadvantage- A gateway is required
Loss of quality of service.
25
CONNECTION TO NATIONAL NETWORKS
Option 4- Country has a different technology.
Advantage- Full autonomy for country.
Disadvantage- A gateway or relay is required
Loss of quality of service.
26
NETWORK MANAGEMENT
Network controlled from small microvax
Microvax is on an X.25 connection to a
switch and can be remote.
Microvax constructs all tables for the
switches and down loads them.
Microvax receives statistics and alarms
from switches and provides reports.
27
MANPOWER
Duties at a switch site-
Ensure lines and modems are working.
Ensure switch is operational.
Installation of lines and modems.
Low level of expertise required
Small fraction of a man.
Duties at the "management centre"
Control of all tables in the switches.
Monitoring of network and any actions
needed to restore services.
Collection of statistics.
High level of expertise required
About one man.
28
Since the last Board of Directors meeting, the Technical Group have held one meeting. This was in Lisbon from 30 November to 1 December 1987. There has also been a great deal of valuable discussion via the EARNTECH distribution list.
At Lisbon there were two principle topics:-
With the termination of IBM support the Group would ideally like EARN supported staff to manage the network and to undertake the tasks previously done by Berthol Pash and others. This is estimated to be a task for two people. A job specification was drawn up and a proposal put to the Executive. I am happy to report that the proposal has been accepted by the Executive and staff will be appointed subject to approval by the Board and approval the subsequent expenditure. The Group also suggested a further equivalent of two posts was needed for specific developments which would be contracted to relevant organisations. This has yet to be progressed.
As finance for the two posts cannot be made available until 1989, the interim management tasks are being undertaken by volunteer effort. We are indebted to Ulric Giese, who will co-ordinate the EARN tables. In addition we are grateful to Eric Thomas, Dominique Dumas, and Dominique Pinse who are looking after NETSERV, LISTSERV, and other items.
Most EARN management software is well developed and can now be more or less frozen.
EARN has been a fairly loose association of nodes. This has led to problems particularly with international connections. For example, there is overloading, there are delays with some types of files, and there are no mechanisms for the control of nodes.
To form a basis for control the Group proposed that EARN (or rather the Executive or Board) should issue "directives" or "recommendations" which are applicable to certain "classes" of nodes. It will be mandatory for EARN directives to be implemented while recommendations remain as advice to sites. The classes of nodes are "international", "transit" (nodes with two or more connection but are not international), and "end". Directives and recommendations will not be issued without appropriate consultation with Board members.
This recommendation has been adopted by the Executive.
A set of directives and recommendations for the control of file traffic was proposed to the Group by Eric Thomas. After some amendments by the Group, the Executive has now requested comments on the proposal from the Board.
Some progress was made with encouraging the use of mail Domain Names. This is as a preparation for the transition to ISO protocols. This will be reconsidered at a later meeting.
Topology and statistics are of great concern. Sensible statements on topology are difficult without statistics. several attempts at statistics gathering have been attempted and a further one was set up.
A proposal to set up small "task groups" was accepted. These will address specific questions and report back to the group. Such groups were set up to consider domain names and statistics. All work would be undertaken electronically.
A permanent group was proposed composed of the "Country Co-ordinators" and a few others. This would be an "electronic" group led by the Network Manager (currently Alain Auroux). The group would be responsible for keeping the network running on a day to day basis. This proposal has been accepted in principle by the Executive.
The Group is not undertaking transition work which is done by other sub-groups set up by the Executive. None the else they take a keen interest and debate the evolving plans.
The technical Group is now operating on a more formal basis as it attempts to address the problems we have. In the past it had a strong educational role. Thus formal proposals are accepted, discussed, and agreed. The Group intend to continue with twice yearly meetings. The next meeting is in Turkey on 15/16 April and the subsequent one in Finland.
EARN has requested comments on the 'File Control Proposal' (appended) with a view to its adoption.
It will be mandatory for RAL to implement 'directives' if the proposal is accepted. RAL is in a position to veto acceptance.
The text below is a draft for submission to EARN if accepted by NTMPC.
The UK is in favour of adopting the EARN 'File Traffic Control Proposal'. All the directives are acceptable in the context of RSCS V2 which the UK will adopt. Most of the recommendations are acceptable.
UK is surprised that there is no recommendation for a transition to RSCS V2. It is the UKs intention to use RSCS V2 for the EARN connections.
UK believes that modifications to source code should only be undertaken where this is the best, or only means of providing a function for which there is an immediate requirement.
2.1.1 Directive Number 1
UK agrees to directive number 1. The multi-streaming directives can only be implemented if IRLEARN and CEARN adopt RSCS V2.
2.1.2 Directive number 2
UK agrees to directive number 2 which requires the automatic restarting of lines.
2.1.3 Directive number 3
UK agrees to directive number 3. The mechanisms for this have yet to be agreed.
2.1.4 Directive number 4
UK agrees to directive number 4.
The UK is unlikely to adopt recommendations which require code changes to RSCS.
2.2.1 Recommendation number 1
The UK has no view.
2.2.2 Recommendation number 2
The UK recognises the problem of the spool filling and will take such steps as are necessary to avoid any problems.
2.2.3 Recommendation number 3
The use of BITSEND/BITRCV is being investigated.
2.2.4 Recommendation number 4
The size limits for files are reasonable but may need adjusting to meet local circumstances.
2.2.5 Recommendation number 5
UK does not intend to authorise remote staff to control RSCS on UKACRL for security reasons. As far as possible local staff will be available for any required actions.
UK is unhappy to control RSCS on any remote sites.
2.2.6 Recommendation number 6
UK is reluctant to remove commands from RSCS. It would be acceptable to make them 'authorised (in the RSCS sense) thus disabling them for general users.
RSCS V2 allows a user to query files originated by him alone on a remote node which removes the need for requesting large amounts of output.
2.2.7 Recommendation number 7
UK would prefer to wait for the RSCS V2.3 facility rather than implement the quiet modification. If this proves unsatisfactory then the use of QUIET RSCSMOD2 will be reconsidered. There is a large "small programming enhancement" which provides a similar facility which would only be implemented with reluctance.
2.2.8 Recommendation number 8
UK will reconsider this recommendation when the RSCS V2 version is available.
2.2.9 Recommendation number 9
UK currently operates with DMT prefix suppressed.
2.2.10 Recommendation number 10
UK does not use JES2.
Dear Manager,
Thank you for making my stay in Turkey and at your hotel such a delightful experience. I shall long remember the friendliness and consideration that your staff showed. Our meetings were a great success helped by your excellent facilities and comfortable surroundings.
This was my first visit to your country and I hope that I will be able to return. You are indeed privileged to live is such a beautiful part of the world with such hospitable people.
Thank you for all the help and assistance you and your staff gave us.
Your sincerely,
Paul Bryant - EARN Technical Director.
The meeting was attended by 20 members.
The meeting split into two parallel sessions for most of the time. The first to review the deliverables and the second to look at the EARN document. The majority (including the chairman, Piete Bovinger, opted to look at the EARN document.
Few documents were received before the meeting which made the review group's work 'interesting'. On the X.29 review, they could only review a list of questions which were to be sent to suppliers. As this was the last meeting before the end of the COSINE definition phase I fear for the quality and quantity of the result of WG4. My own view is that as far as specifying a COSINE X.25 network is concerned, WG4 has been a disappointment. Few documents have been produced and few of these have been of a reasonable quality.
A document of some interest was the survey of switches. It was done by the Dutch PTT. It only covered switches available in Holland and the UK. Also, we understood that, the PTT had 'selected' switches for review. Alcatel, IBM, and Telepak were absent. Also, many switches suggested for inclusion by UK had been omitted. This was a useful but incomplete document. It was useful from the view of what it told us about the Dutch PTT rather from the merit of the survey,
A document, not a COSINE deliverable, from the UK GOSIP group was very useful as it specified a functional standard for the X.25 network. I believe we (EARN) should adopt this for our switch specification if possible.
There were no other documents of interest to us.
A lot of time was spent on the EARN document which came through more or less unscathed. The address proposal was accepted after much discussion. The meeting did not discuss topology. The chairman kept stating that as there were possibilities of a managed network, topology was not necessarily important. No comments were made on the use of Northern Telecom switches, again the chairman noted that this was not relevant if we had a managed network. It was felt that the switch specification should be replaced by a DCE specification as this would be applicable for a managed network. In fact this was done with no real change to the substance of the document. It was noted that it was very difficult to offer a DCE specification to a switch manufacturer as performance depended on the inter-switch lines which are outside his control.
A long discussion was held on NSAPS. This was considered in the Perugia paper but was not considered by the OSI team. It is important. The conclusion reached was that the ISO DCC scheme should be used in preference to the X.121 scheme. This is because we cannot guarantee the orthogonality of the EARN and PTT X.25 DTE address scheme. If the X.121 scheme is adopted then a strenuous effort should be made to either obtain an EARN DNIC or a portion of the public address space in each country. Both of these, in my opinion, would take a long time to obtain. The ISO DCC scheme does imply directory services will have to be provided which is a subject we should study.
It is absolutely clear that the Dutch PTT member and the Dutch chairman are determined to press for a managed network. The veneer of pretence that a network provided by the community would be a possibility slipped slowly away as the meeting progressed. However, this issue was not discussed and your worthy delegate merely noted the comments coming forth. It is clear to me that WGA has a hidden agenda and Northern Telecom and a community managed network are not on it.
The meeting considered the capacity of the proposed EARN infrastructure. They felt that the switch specification at 800 packets per second was a bit low and were talking about 3000 packets per second as being better. This takes into account 2M lines. They also though that the figures should be revised upwards in the light of the network being used by other communities such as EUNET.
Since the last meeting there have been no changes to the gateway at Rutherford laboratory apart from the updating of tables.
A new version of the gateway code is now being mounted. The principle change is that addresses will be tried in all possible domain orders in an attempt to deliver mail. All mail that is currently delivered correctly will still succeed. Looking at the rejected mail, it appears that the principle problem encountered by users is the confusion over domain ordering.
The new gateway code will also allow EARN distribution lists to be used within JANET. The mail distribution system is called LISTSERV and has been operational at Rutherford for some time but only from the EARN side of the gateway. LISTSERV allows users to manage, join, leave, and interrogate various distribution lists. These facilities will be extended to JANET users. LISTSERV is operational on many EARN nodes and the set of LISTSERVs is 'linked' in an attempt to reduce network traffic.
The NRS name UK.AC.EARN-RELAY has now been defined. The old version, UK.AC.RL.EARN will be withdrawn when there is evidence that it is no longer in use. That will probably take six months or so.
The NRS name UK.AC.RL.MAIL has been defined which is provided for access to users on the Rutherford IBM computer and the mail distribution facilities. Although the registration is the same as for UK.AC.EARN-RELAY it is provided principally to avoid confusion for the use of the distribution facilities for UK use only.
A new version of the EARN documentation is soon to be distributed which defines the changes which are now being made.
The traffic level have continues to rise as shown in the attached graph. As before, the traffic levels are erratic but have shown levels approaching 2G bytes a month. As yet there are no indications that the connection to CERN is overloaded.
There have been overloading problems at Montpellier since all traffic to the USA passes through that site. There has been a proposal to have a further line but this was rejected by the Board of Directors on financial grounds. Better operations procedures appear to have reduced the problem.
A number of file servers have been mounted. IRLEARN (Ireland) now have a MAC server from which useful software can be obtained. PC servers have been mounted.
So far no tests have been run to see if such servers can be used via the gateway.
Unfortunately it has not been possible to recruit a further staff member to look after the gateway. These tasks have been undertaken by various people. Thus, although the gateway support has not been as good as it could be, a reliable service has been maintained. The main effect has been the delay in producing new gateway code.
Rutherford now has permission to recruit for this post and interviews are to be held shortly.
JNT is now responsible for the EARN financing.
The Board of Directors, at their meeting in Izmir Turkey, decided to follow a similar funding model to 1988. Thus the UK will be responsible for the cost of the line to CERN and a portion of EARN's other costs.
As the IBM support for the EARN office ends at the end of 1988 this cost will now have to be born by the membership as well as the cost of the line to the USA. This year these costs were distributed according to the 'RARE keys' which are related to a country's GNP. The Board has decided to distributed the costs in future years according to GNP. However, since this would cause a steep increase for some countries it has been decided to fix the fee half way between the RARE key and GNP for this year. The final costs are now been worked out for approval.
JNT approval for the subscription will be sought when the definitive level is known.
The total funding required is about 2M ECUs of which 1.8M ECUs is for lines.
EARN has included funds for 4 technical staff members in its budget. It is felt that it is no longer satisfactory for critical network maintenance to be undertaken by voluntary effort.
There have been changes to the topology of EARN to avoid the volume tariff in Germany. These are now complete and appear to have had no adverse effect. Apart from a 64K line between Germany and France, no further changes are currently being contemplated.
No new countries have been connected although applications from Cyprus, Algeria, and Yugoslavia have been approved. Applications or expected applications from Morocco, Tunisia, China, India, Hungary, Poland, and Russia are in the 'pipeline'. It is unlikely that connections to eastern Europe will materialise in the near future.
Subsequent to the Perugia meeting a number of exciting developments have taken place.
DEC has agreed to fund much of the transition to ISO protocols. They funded a series of meetings to decide of further details. It has now been decided to use NJE over ISO session as the way of preserving existing services. This will be implemented on Micro VAX computers to form a gateway between the tradition EARN protocols and ones capable of operating over the proposed X.25 infrastructure. It is expected that a similar project will be undertaken for IBM computers.
Gateways between X.400 and the RFC822 mail service will be produced and this will also be available on a Micro VAX.
It is hoped that DEC will provide a VAX gateway computer in all EARN countries.
Northern Telecom have offered to provide EARN with four switches. Further design work on the X.25 infrastructure has come up with a topology proposal which shown Holland, Stockholm, CERN, and Montpellier hosting switches. This 'square' will be interconnected by 64K lines. The Northern Telecom switches have an impressive performance of 3500 packets per second.
IBM is interested in funding EARN's further development in connection with their super computer initiative. Discussions are taking place.
Discussions are now being held with EUNET, COSINE, and possible others to see if a common X.25 infrastructure can be provided. These discussions will be complete in a few months when implementation can be expected.
A successful meeting was held in Izmir Turkey. This included a meeting of the EARN Technical Group, the EARN Executive, and the EARN Board of Directors.
The principle topics of discussion were finance and transition which have already been covered.
EARN continues to thrive and expand. EARN still prides itself that it is providing a service to the users and does its utmost to improve this service at an economic price.
The transition to ISO protocols is now entering an exciting phase where we can expect some radical developments this year. The drawing together of the various networking communities in a common infrastructure is an exciting concept which is fully supported by EARN. None the less, EARN is determined that in this transition the service to the users will be maintained.
Unfortunately I did not find time to speak to Brain Day on the CCD ethernet paper. I have a few comments.
The strategy we are adopting is not very clear. I assume that we are putting in a "yellow cable" to which we connect "essential services" and fan out units.
If this is so then the "essential services" to my mind would be those which are required by a number of agents. For example - the bridge to the backbone, the gateway to the X.25 network, and monitoring equipment. It is debatable whether the computing services themselves would be in this category - I would say they would not be. This strategy would cause the yellow cable to be very stable with changes being very infrequent. All changes would be to the thin ethernet cables on the fan out units.
If this is the strategy then the computers, PADs and so on would be on thin ethernet.
Thus, the yellow cable would follow a path which would visit the best locations for fan out units and the essential services. This would imply R30 area, telecoms area, and the paper store by the front door. This would differ from the proposal from Brian in that the computer rooms would be served by one or more thin ethernets from the fan out unit in telecoms area. This strategy would seem to give more flexibility in the computer rooms as thin cable can be installed and re-installed more easily than thick cable.
Perhaps it may be worth asking whether we could dispense with yellow cable all together and have a three point fibre optic star based in telecom's area.
A possible difficulty is that some equipment may not be attachable to thin ethernet. As far as I know this may only be the Auscom. (I should add that I have been having discussions with Gunadhi's company and developments are afoot that may solve this problem).
The exact runs and numbers of thin ethernet cables is difficult to predict but this can be expected to evolve as time progresses.
The above strategy should reduce the length of the yellow cable and thus its cost. It would also allow segments of cable to be isolated for testing if required.
I am not sure that a test ethernet is required. If such is to be useful then it must visit all the interesting machines which means that it must be a duplicates of the main ethernet. In addition for several machines we only have a single connection and thus any testing will cause disruption in any case. I suspect that most testing will be for new machines or for small machines (PCs) and in these cases a short length of thin cable could well be all that is needed. I am not sure that we will need a bridge between such a cable and the main service or any connection. If such were needed I suspect that the spare bridge should be put into service and all we need to do is to make cable provision for it.
I agree that we need a spare bridge. I think a single spare for the whole site is adequate as they appear to be very robust. This is debatable. I agree with the need for a spare fan-out unit. We certainly need a spare fibre optic transceiver for the site, I am not sure that we need for for our division. Again, these seem reliable- again debatable.
I fully agree with the need for documentation and training as well as the rules.
The last sentence is ambiguous. Does it mean 0.5MY expended in the first half year (one man full time for half a year) or does it mean a rate of 0.5 MY for half a year (one man half time for half a year)? The full time interpretation seems rather high. I assume that the higher rate at the beginning allows for training.
Finally, I am reluctant to provide a service until all the major components are in place. We currently lack the gateway to the wide area. I might add that since this will be a vital component it may well be advisable to have a spare. An alternative strategy could be to have two gateways in service - one to JANET and one to the local X.25 network with provision for re-routing traffic in the event of one failing.
EARN provides services to members of the academic community. It supports a small number of 'associate members' which are commercial organisations with connections with the academic community although they may only use the network for non commercial purposes.
No survey has been taken of the community. Some information can be gained from traffic observations and user comment.
The use is predominantly by scientists and engineers although those in the humanities are known to use EARN. The largest groups of users are those involved in the large international collaborations. Of these high energy physics predominates and the traffic flows between CERN and the collaborating groups in the universities. These is also a traffic flow to the American high energy sites who collaborate closely with their European counterparts. The astrophysics groups are active but smaller. They collaborate on a large number of projects related to astronomy and satellite work. The third largest group is the computing community who use EARN for a wide range of activities which include the maintenance and development of EARN to the development of protocols. There is a large amount of mail traffic of an occasional or social nature such as fixing visits or the 'real coke drinkers' list.
The majority of users are in institutes of higher education with substantial computing resources of at least a VAX or IBM 4000 series. This is not surprising as a connection to EARN requires a computer which is capable of running the NJE protocols. Thus the smaller institutes are poorly represented.
The population is estimated to be 50000. This excludes the large number who use EARN from the USA. The population changes from month to month suggesting that many users are seasonal.
EARN provides two basic services
There are services which are built on the basic ones.
The store and forward nature of the network forces all traffic not to be time critical. The relatively narrow bandwidth connections causes quite long delays although they are capable in most cases of keeping up with the traffic load. It is certain that an interactive service on the existing lines would be inadequate on the basis of there long periods of saturation.
Mail is more popular than would be expected as only mail can pass through gateways. Thus, files and documents, which are in principle file transfers, are 'wrapped' in mail.
Leaving aside transition few changes are envisaged. The principle reason is financial restrictions which prevent the installation of wider band width connections or alternative routes to improve capacity and reliability.
Although changes are not envisaged the current services are expanding. The traffic shows a growth of 100% per year. This in unlikely to continue as line saturation will occur in several places. This manifests itself in traffic delays and in some cases spools fill up with unfortunate consequences. Palliatives are being implemented, such as better scheduling algorithms. The use of LISTSERV also helps to reduce traffic but on the other hand attracts traffic by its quality and usefulness.
Accurate statistics are not available.
The principle sources and sinks are CERN and Germany. The line topology does not reflect the traffic pattern because of the differing line tariffs between the PTTs. The congested lines are between Germany, CERN and Montpellier. The Montpellier congestion is because of the connection from there to the USA.
The traffic figures are:
To be supplied.
The principal requirement is to preserve and if possible enhance the service to the user.
This implies that sites with current EARN connections must be able to continue to use these connections with out significant change. Thus, the NJE protocols must be carried across any X.25 infrastructure transparently.
The store and forward nature of the network suggests that current services could be maintained by an infrastructure based on 9.6 lines. In fact the better topology should improve performance slightly although the small X.25 block size will have the opposite effect. None the less, the networks continuing traffic growth indicate that a number of broader band links would be desirable.
EARN currently has no plans for services other than NJE and the emerging X.400 and FTAM services. These new services are expected to eventually replace the NJE and RFC822 mail services and in principle should not impose a higher traffic level. However, the greater population capable of using the network would cause a traffic increase. There are no estimates of the population increase to be expected by offering ISO protocol services.
An X.25 network is capable of carrying a number of services not currently provided. The one of greatest interest is interactive. There may also be pressure to allow DECNET, SNA, and TCP/IP services. The desirability of allowing these is being studied.
EARN is financed (or soon will be) by the community. This amounts to 1.5M ECU for lines and 0.5M ECU for management, staff and the line to the USA.
It is clear that the community will oppose with vigour any suggestion of a volume charge.
It is clear that the community is unwilling to pay much more for the current service. Indeed they would expect to receive a better service at a lower cost in line with improvements in technology and an expected reduction in PTT tariffs.
Thus, any major cost of a capital or recurrent nature needed for a transition must be born by a benefactor, lead to savings, or from a larger user base.
EARN is obtaining support for its migration. Any collaboration may require the collaborators to bring further resources to the project.
Currently EARN controls the international infrastructure. The control of EARN within a country is the responsibility of the EARN organisation within a country within the EARN statutes and code of conduct.
It is envisaged that an EARN ISO network would follow a similar management structure although the rules for the use of the infrastructure and gateways will be stricter for technical reasons.
If EARN collaborates with others to provide any shared infrastructure then some form of shared management would be appropriate. Until the nature of such a collaboration is known it would be premature to define the nature of the EARN management involvement. EARN would insist that in any collaboration the service to the EARN users would be maintained.
EARN has defined an infrastructure suitable for its own uses. A collaboration with others may well modify this proposal.
It is vital that the infrastructure is well managed.
The current plan suggests two or three equivalent men to manage and operate the infrastructure. This estimate is supported by Northern Telecom. This assumes a single, or very small number of lines into a country and therefore a stable high capacity low connection network. A couple of man months is needed at each switch site plus the operation of the management centre. If EARN manages some X.25 networking within any countries then further effort will be needed but this is not possible to estimate until a proposal is received. The effort would probably be supported by the country.
The management may be undertaken by a PTT, another supplier, or the community. The contractor must be competent and have a suitable contractual and technical relationship with the community.
The figure of three men gives an indication of the costs expected.
It is essential that the supplier is properly controlled by the community rather than supplying the service as he sees fit.
EARN has received an offer of free switches from Northern Telecom and IBM. This would be maintained free of charge for at least three years after installation. Evaluation shows both equipments could meet the EARN requirement although the Northern Telecom offering has greater potential.
There are several other suitable suppliers but these would all involve expenditure.
EARN is intending to provide relays between NJE/bisynch and NJE/OSI-session in each country. These will allow the continuity of service EARN is committed to. These relays will be unlikely to be part of any collaboration.
Any further equipment in a country will normally be the countries responsibility.
Dear Jean-Claude,
Here is my claim for my travel with respect to the EARN EXEC and WPA committees of May 9/10 in Paris and Amsterdam. The CEC is paying for the Amsterdam meeting and to split the costs they are paying the subsistence.
Air fare 238.60 UKL
Please make cheque payable to 'Rutherford Appleton Laboratory'.
Thank you.
With best wishes,
Paul Bryant.
I went to a course at NT on Thursday/Friday 19/20 May 1988.
This was not a very good course as it was designed for those upgrading from old NT equipment to new. None the less I had a chance to discuss addressing. This revealed that we have a number of problems which need exploring.
The NT switch is really designed for use in a PTT and takes the view that it will provide DTE connections from within a country and have an X.75 connection to others. Thus the address structure within a network is defined for the network as a whole. For example 4 digits are dedicated to for the DNIC, the rest of the digits (up to 10) is split for all DTEs between the address and subaddress with the restriction that the address has a minimum length of 5 digits. The network may have a number of DNICS within it.
External DNICs are defined and all addresses starting with that DNIC are sent to a specific X.75 port.
So why is this a problem?
It is clear that the sort of network we want has never been attempted by NT before and they were somewhat bemused. However they were somewhat fascinated and not critical. The expert were not all that conversant with the options available.
My conclusion is that we will have severe problems connecting to other networks. Ideally, connections should be via X.75 but in practice this seems unlikely. We can try another manufacturer (is IBM any better?), attempt to get NT to make changes (unlikely), change our address structure (we will regret it!), or just batter ahead.
It is interesting to note that JANET is really a set of 11 networks (one round each switch) interconnected by X.25 which allows a lot of flexibility in addressing. NT, in contrast, is a single network demanding the same address structure for the whole network.
Any comments? Paul.
Dear Nick,
I enclose my claim for the WPA meeting. As I mentioned in Switzerland I am sharing the claim between EARN and WPA as there was an Executive meeting the day before in Paris. It would seem simplest if EARN picks up the travel and you pick up the subsistence. I have a funny feeling that a more complex split would send our accounts departments round the bend. I hope this is OK. I also enclose air ticket as proof of stay.
I look forward to the next meeting.
With best wishes,
Paul Bryant.
It is proposed to issue the new EARN documentation.
This has already been circulated but there is a further change as a result of comment. A new copy is attached for your convenience. It has become clear that if mail traffic to RL.IB were forced to go via the mail system then this would upset some users. It is also clear that if mail destined for distribution lists were to go to EARN-RELAY that this would be confusing. It is thus proposed to register a new name RL.MAIL (long form RUTHERFORD.MAIL) which would remove the confusion and allow progress to be made with distribution lists. It is not yet appropriate to move all the distribution lists from the GEC computer before further experience has been obtained. When this is obtained then RL will be moved from the GEC to the IBM computer. MAIL will be registered in 'forward' direction only and the reverse will be to EARN-RELAY.
NTMPC is asked to approve the new documentation.
It appears that in the UK an X.25 network is a 'value added network'.
It appears that an X.25 VAN must have its DNIC approved by OFTEL. This seems to imply that a VAN can apply for a DNIC.
I do not know if similar regulations exist in other countries. However it may be appropriate to try and apply for a DNIC in each EARN country. I guess this should be done in conjunction with the national network organisations. This may be an interesting exercise!
It is now becoming urgent that we address the NSAP problem. There seems no argument against the use of the ISO DCC scheme. In principle this requires a directory service. In the short term, as we shall have very few DTEs, we can build in the NSAPs. It is unclear what Joiners and IBM will be doing with NSAPs. I very much suspect that they will be using the X.121 scheme as this is recommended in DEC documents.
The ISO DCC scheme expects that NSAP schemes are obtained from the appropriate authority in each country. In the UK JANET has already registered with OFTEL but there now seems to be some confrontation between DTI and OFTEL as to who is the right authority. The situation in the other countries is not known to me but we need to find this out.
I think we are getting round to needing to get something like the programme committee into action to resolve these issues.
25 people attended. CEC were represented on the second day by a new face - J Pato. P Bovenga of RARE attended the last day. MDNS was represented by G Sibering and the "One Stop Shopping Group" by T Huebner.
A large number of documents were circulated from the 4 working groups and from other sources. Rare and EARN documents were there. The meeting was very much a working meeting with parallel sessions and moved the subject along rather than came to a set of definitive results.
Comments on EARN came up from time to time and EARN/RARE etc are treated with interest rather than anything else. T Huebner questioned in private whether EARN had any intention to migrate. He had a copy of "Quo Vadis EARN" by Michael Hebgen which I do not have.
A chart of the facilities of the European networks is being made. This is very similar to the charts other groups are making. The aim is to publish the document. I commented that the users would like to see the differences disappear.
Quality and performance would be discussed next time. It is difficult to measure performance but some administrations are making measurements with PCs so that it can be measured as the user see it. Some metrics were needed.
Work is needed on aligning diagnostic codes. Users need to know what they really mean. A user guide is needed.
A survey of number plans is taking place. There was long discussion on abbreviated numbers which appear to be treated differently in different counties. There were 13 replies so far.
I was surprised that PTTs were not well informed on other PTTs.
A document of high speed services suggests that 2 meg circuits could be used. There were problems with window sizes which suggests that good percentage use may be difficult.
A survey suggests that X.25 services will be with us for 10 years at least. I felt that 10 years really meant that no one really had much idea. It was unclear what effect ISDN would have.
There was an interesting discussion on differential tariffication for various traffic (bulk and interactive). There are difficulties as to how to indicate such traffic as there was no suitable facility. The thought was to use a different DNIC. This seems a prostitution of the use of DNICs but there is no other feasible idea.
A user guide is now complete and will be published. It contains entries for all European X.28 offerings. It will hopefully be on line.
The main problem is how to charge for an outgoing call. Some administrations want to forget it but BT is unable to as this would be cross subsidisation,
There was also some discussion on universal NUIs.
There is a proposal to put the Eurodate Foundation Book on line. A number of options were to hand including the ECHO machine. The information will need adapting for a data base.
There is an intention to introduce OSS for X.25. However there was some concern as to what the product is. T Huebner said that OSS could only sell products and did not invent them. It seems that administrations would sell the products not the OSS group. One view was that OSS was more of a single point of contact to PTTs for information rather than actual services.
MDNS is expected to be a company. The products they will sell are unclear. Again it seems that the products will be sold via administrations. They see EARN (RARE/COSINE?) as a possible pilot group for them. They seem to have little idea of what we want in detail. A meeting in Rome will sort out the MDNS company for a final agreement in the Autumn.
Myself, OSS, MDNS, and the chairman discussed the needs of groups such as RARE, COSINE, and EARN. It was decided to recommend that MDNS was contracted to study the requirements for presentation to the group. The academic groups would be invited to participate.