25-Apr-80 23:38:02-PST,57788;000000000001 Mail-From: OFFICE-1 rcvd at 25 April 1980 22:26-PDT Date: 25 April 1980 22:26-PDT From: Stefferud at OFFICE-1 Subject: MSGGROUP#1520 IFIP WG6.5 Montreal Workshop Report To: MSGGROUP at USC-ECL 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT ABSTRACT: This report of the March 1979 IFIP Working Group 6.5 Workshop, held on March 1-2, 1979 at the Hyatt Regence Hotel in Montreal, Canada has been prepared for IFIP by Einar Stefferud, Ronald Uhlig, David Farber and Ian Cunningham. It does not necessarily represent the positions of any of the attendees or their institutions, but it does represent an attempt to summarize and synthesize an integrated description of the issues that were identified and discussed in the workshop, and which are involved with evolution and development of international computer message services. IFIP WG 6.5 is chartered to concentrate on standards for data structures, addressing, and higher level protocols to effect international computer mediated message services. Page 0 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT TABLE OF CONTENTS Page Number INTRODUCTION....................................................2 SUBGROUP ISSUE LISTS............................................4 USER ENVIRONMENT.............................................4 SYSTEM ENVIRONMENT...........................................4 POLICY ENVIRONMENT...........................................4 QUESTIONS RAISED................................................5 GENERAL CMS ENVIRONMENT DISCUSSION..............................7 CMS USER ENVIRONMENT DISCUSSION................................10 PRIMARY USER SERVICES.......................................10 ANCILLARY USER SERVICES.....................................11 USER TERMINALS AND COMPUTERS................................11 CMS USER LANGUAGES..........................................12 CMS USER STANDARDS..........................................12 LONG TERM USER EXPECTATIONS.................................13 USER ECONOMICS - COSTS VS BENEFITS..........................13 CMS SYSTEM ENVIRONMENT DISCUSSION..............................15 ARCHITECTURE................................................15 PROTOCOLS...................................................16 SERVICES....................................................17 TRANSFER.................................................17 MEDIATION................................................18 STORAGE..................................................19 RELIABILITY..............................................19 SECURITY & PRIVACY.......................................19 STANDARDS...................................................20 ECONOMICS - COSTS VS PRICES.................................20 CMS POLICY ENVIRONMENT DISCUSSION..............................21 REGULATION..................................................21 TARIFFS.....................................................21 USE OF EXISTING FACILITIES..................................22 SPECIFICATION OF FUTURE FACILITIES..........................22 ECONOMICS - TRADE AND COMMERCE..............................22 APPENDICES.....................................................24 APPENDIX A: ATTENDEES......................................24 APPENDIX B: SUBGROUP PARTICIPANTS..........................25 APPENDIX C: ADDRESSES OF IFIP WG 6.5 CHAIR PEOPLE..........26 Page 1 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT I. INTRODUCTION The IFIP Working Group on "International Computer Message Services" (IFIP WG 6.5) held a Workshop Meeting in Montreal, Canada, on March 1-2, 1979. Ronald Uhlig is Chairman of IFIP WG 6.5 and served as Chairman of the workshop. David Farber served as Program Chairman for the Workshop. The following statement of "AIMS & PURPOSE" is taken directly from the IFIP WG 6.5 Charter: "Computer mediated messaging is a rapidly emerging service area. Messages may be processed, stored, and transmitted between users potentially within the jurisdiction of separate carriers, computer systems, and/or computer networks. Technical, economical, and political issues need to be resolved in order for a viable international computer message service to develop. This Working Group will concentrate on standards for data structures, addressing, and higher level protocols to effect international computer mediated message services. Such a service could impact existing international postal and communication agreements, and economics of the world-wide communication system. These issues will also be studied in conjunction with appropriate IFIP 'Working Groups'." The purpose of the IFIP WG 6.5 Montreal Workshop was to identify and discuss issues related to Computer Message Systems and to set the stage for future work. The following sessions were held: Introduction (R. Uhlig) Scope and Models (I. Cunningham) Functions and Facilities (J. Pickens & D. Davies) Economics (E. Stefferud) Addressing & Protocols (D. Crocker, J. Postel, N. Naffah, Mike Schroeder) Regulatory Issues (G. Edwards, Bill McLean, P. Walker, R. Pye) Experiments & Experiences (D. Farber) Discussion & Organization of Work (Attendees) Page 2 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT At the end of the meeting it was decided to organize activities of Working Group 6.5 for the next year into three broad categories (USER ENVIRONMENT, SYSTEM ENVIRONMENT, and POLICY ENVIRONMENT) for consideration by subgroups. GENERAL ENVIRONMENT issues are to be considered by the whole Working Group 6.5. Initial lists of issues for each subgroup to consider were identified and three subgroups were formed. Subgroups and International Co-Chair people are: USER ENVIRONMENT: Donald Davies - Europe Elizabeth Feinler - North America SYSTEM ENVIRONMENT: Najah Naffah - Europe Ian Cunningham - North America POLICY ENVIRONMENT: Gwen Edwards - North America Roger Pye - Europe Page 3 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT II. SUBGROUP ISSUE LISTS USER ENVIRONMENT Primary services Ancillary services Terminals and Computers Languages Standards Economics - Prices vs Benefits SYSTEM ENVIRONMENT Architecture Protocols Services Transfer Mediation Storage Reliability Security & privacy Standards Economics - Costs vs Prices POLICY ENVIRONMENT Regulation Tariffs Use of existing facilities Specification of future facilities Economics - Trade Page 4 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT III. QUESTIONS RAISED A collection of questions was complied from the discussions and reviewed during the final session. The questions did not always fall easily into exclusive subgroup domains. They are listed below with tags appended to indicate which committees might be primarily concerned. 1. What is, or is not, a "message?" (USER) (SYSTEM) 2. What group(s) should (will) supply technical standards for CMS? (POLICY) 3. In what ways do privacy techniques conflict with: (1) Transborder data flow policies? (POLICY) (2) Transport mechanisms? (SYSTEM) (3) Economic feasibility? (USER) 4. Should CMS be a regulated service? (POLICY) (1) Should it be universal? (2) Who should supply the service? (3) What should be included in the service? 5. What should be the interface between message composition/reading software and the transport mechanism? (SYSTEM/USER) 6. What architecture should be assumed for CMS? (SYSTEM) 7. What methods are available for provision of privacy, authentication, and signature services? (SYSTEM) 8. What vocabulary should be used in discussing CMS? (USER) (POLICY) 9. What are "junk electronic messages" and how can we deal with them? (USER) 10. How can "naming problems" be minimized? (SYSTEM) 11. What message structure and contents are needed for transport? (SYSTEM) 12. How should different languages and alphabets be dealt with? (SYSTEM) 13. What user composition/reading/filing services are desired? (USER) Page 5 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT 14. What is the relationship among: (POLICY) (1) International Postal Services, (2) Electronic Funds Transfer, (3) Common Carrier Services, and (4) existing or planned communication services? 15. What is an office? (USER) (1) What is an automated office? (2) How does CMS fit in? 16. What are the business needs for CMS? How will these affect: (1) Development of services? (USER) (2) Development of standards? (SYSTEM) The remainder of this report is organized according to the chosen subgroup categories and initial issue lists. Discussions held during the workshop are reported under the appropriate subtitles. Page 6 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT IV. GENERAL CMS ENVIRONMENT DISCUSSION The appearance of computer message services is a natural outgrowth of technology advances and the convergence of communication and computer technologies. The new services arose out of the need in early time-sharing computer systems to communicate among users and operators in the course of interactive computing sessions. Since their inception computer message systems have expanded beyond the bounds of single computers to become more ubiquitous. In various current networks of computers, large numbers (thousands) of individuals and agencies are able to communicate among themselves via message exchange using many different computers and terminals in the process. Computer Message Services use communication mechanisms to deliver message texts from senders to receivers. When the receiver is using a different computer than the sender, then an inter-process communication path must be set up to allow the two computers to accomplish the message text transfer between them. Sometimes the sender and receiver collaborate directly to establish the communication path between their separate computer file systems. In other cases, certified computer system programs with special access privileges are used to copy a designated message text from the sender's private file space to the receiver's private file space without direct involvement of either party. In at least one current computer network, the message text transfer mechanism is manifested as a file transfer process that operates according to a file transfer protocol for interprocess communication. In general, any communication channel between sender and receiver can provide this functional capability if their computers and terminals can be made to use it. The end-user product of Computer Message Service is typewritten communication between individuals. This end result is much the same as with traditional record carrier message services, except that delivery consists of placing electronic copies into computer files which allows the power of a computer to be applied to them. The receivers then have the option to display and process their messages in various ways. Delivery of an electronic copy in place of a printed paper copy is what distinguishes computer message systems from traditional record carrier systems. Page 7 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT This difference may seem small and perhaps inconsequential, but it appears to have profound impacts on the future of record communication systems, and on the ways people might use message services in the conduct of business. Much is heard about office automation which involves use of electronic filing systems and electronic computing devices in offices to store and process information. The natural consequence of these office automation efforts will be to interconnect the coming proliferation of computing devices for message communications. The prospect of interconnection of large numbers of computers for international message communication must pose a threat of revolution for established record carriers who are faced with finding ways to deal with an uncertain future. They must retain a place for themselves in the future message service marketplace, and they must protect their current capital investment. It seems clear that the advent of computer networks with their high speed data links and flexible transport protocol capabilities must threaten to significantly impact record communication systems. Thus the problems to be solved are far greater than those of new system design or interactive user interface design. We must find an acceptable evolutionary path that will continually advance existing message system facilities into the future to help provide computer message services on a broad international scale. In summary, we have so far identified the general environment in which fledgling computer message services must develop and mature. It is to the task of searching for and finding satisfactory ways that IFIP WG 6.5 is dedicated. Among the problems, we find wide disparities between regulatory policies in different countries, difficult interfaces among old and new communication networks, and incomplete understanding of the required functional behavior of message processing systems. All of which must be dealt with in the face of dramatic advances in converging computer and communication technologies. To help understand the issues and develop solutions, we need models to provide frameworks for our discussions. We need models for user environments, for service definition, for systems architectures, for communication paradigms and protocols, for economics, for tariffs and regulation, for international relations and trade, for interactions among all these, and for many other aspects of the computer message services domain. Models are needed to simplify descriptions, partition problems, develop terminology, and identify needed capabilities. Page 8 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT At this general level of discussion, our tentative CMS model consists of people who enter information into computers for processing, and who then want to transfer their information to others for communication purposes. The model includes the cases where message texts are to be moved between different computers via some communication facilities, and between computer file systems that are controlled by different individuals or agencies. It includes mechanisms that can copy text from a sender's file space to a receiver's file space, either under joint simultaneous control of the sender/receiver pair, or under control of a third party delivery agent which picks up posted message texts and delivers copies of them to receivers by correctly interpreting attached addresses for receivers' electronic "mailboxes." Our tentative CMS model assumes that the senders and receivers will have the option to apply any available computation to files in their possession. Clearly this tentative general CMS model needs extensive expansion and refinement. It will be the work of the IFIP WG 6.5 subgroups to refine our tentative model, develop subsidiary models, and postulate solutions to the problems. USER ENVIRONMENT discussions will focus on the actions and processes invoked by users in the context of entering, editing, sending, receiving, filing, retrieving, displaying and processing messages with their computers. User Environment models are generally based on interactive computer system concepts. Productivity and end-use functionality issues will dominate. SYSTEM ENVIRONMENT discussions will focus on the design and construction of systems and facilities to meet user functional requirements, and will focus on issues of interfacing to, and utilizing, existing facilities. System Environment models are generally based on computer networking concepts. Architectural issues and standards will dominate. POLICY ENVIRONMENT discussions will focus on issues of transnational interfaces, network interfaces, regulation, tariffs, and transition from existing facilities to new more functional facilities. Policy Environment models are generally based on international relations concepts and on economic marketplace concepts. Political issues must be faced here along with the technical issues. Page 9 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT V. CMS USER ENVIRONMENT DISCUSSION User Environment models are generally based on interactive computer system concepts. The user is expected to interact with CMS systems through interactive terminals connected to computers which are in turn connected to message transport networks. There is no particular size requirement for CMS user system computers, other than the capacity required to meet CMS transport interface standards. A. PRIMARY USER SERVICES Primary user services will include message ORIGINATION, SENDING, TRANSPORT, RECEIVING, FILING, RETRIEVAL, and other PROCESSING by means of user accessible systems of various kinds, located in diverse places, and connected through various communication networks. Technical design and access constraints should be minimized. Choices among user systems and services should be left open to consumer choice as may be governed by national, local, and private interests. ORIGINATION includes all forms of data entry, editing, and formatting. From the functional point of view, origination should be allowed for virtually any facility with the capacity to properly interface to CMS transport facilities. Large and small computers, word processors, TWX and TELEX terminals should somehow be useable for origination. There should be no particular limit or constraint on the functionality or variety of origination systems, as long as they comply with established CMS standards and protocols. SENDING, TRANSPORT, and RECEIVING will require accessible, reliable, secure, and authenticated delivery systems. Third party delivery agents will need to be certified in these regards, and direct collaboration between send/receive pairs will require verification of identities. Technological constraints on access to transport facilities should be minimized, and access to CMS systems should be as ubiquitous as national and local policy might allow. FILING, RETRIEVAL and other PROCESSING, as with ORIGINATION should be possible for virtually any facility that is able to properly interface to CMS transport facilities. Messages should be able to flow into personal correspondence files, transaction processing systems, institutional data banks, or any other kind of information processing systems that the receiver may choose to use. TWX and TELEX printers must be accommodated. Page 10 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT Over all, it will be desirable in the user environment to integrate all these services under a single operating system. Users will need to be able to send and receive from the same system. that they use for ORIGINATION, FILING, RETRIEVAL and other PROCESSING. Indeed, the capability to send, receive, and process messages may become a requirement for all computer systems. B. ANCILLARY USER SERVICES Ancillary CMS services include a variety of things such as name and address directories, mailing assistance, accounting, billing, etc. Address verification and routing determination will be valuable services in the complex future world of interconnected networks, as will charge determination and auditable billing for prepaid and postpaid deliveries. Ancillary services should be available to users through their interactive terminal systems, so that they might verify addresses as they are attached to an outgoing message, or obtain cost information before posting a message for delivery. The entire user environment is expected to be based on interactive computer systems with communication network connections. It is natural to expect these systems to use their network connections for access to ancillary services. C. USER TERMINALS AND COMPUTERS In the General CMS User Model, terminals connected to computers will provide the primary user interface. Terminals will be used to enter message contents into computer files, and terminals will be used to examine received messages by selecting them from their "mailbox" file for display, or for other processing. The computer will serve as a user controllable message text buffer between the terminal and the message transport facilities. Users may enter messages into computers and edit them there before handing them over to the transport mechanisms. Received messages may be delivered into computer message files for later retrieval and examination. Sending and receiving need not require the active presence of any users or their terminals since the computer operating system can interface to the message transport network on their behalf. Page 11 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT None-the-less, it is not intended from the user environment perspective to force delivery to be made to a computer file on behalf of a possibly absent receiver. Delivery to CMS network connected terminals which print received messages on paper should be acceptable on some regular basis. Conversely, it is required that delivery to an absent receiver's files is also a regularly available delivery process. D. CMS USER LANGUAGES The expected user population is extremely large and varied. It would seem important to be able to cater to the full variety of applications and user disciplines. Given the expected size of the market, there should be no need to constrain users to any particular set of CMS user interface languages. The user environment will be a consumer oriented retail market environment where taste and style will be important. User interface languages are important aspects of style, and user interface style can be a vital factor in operational performance. For every discipline there are special languages. Witness the variety of hand calculator "languages" and recognize that CMS users will exhibit an even broader range of applications, even though they will be using their different user interface languages to send and receive messages though a common CMS network. E. CMS USER STANDARDS There are two main needs for standards in the user environment, even though users tend to abhor standards. One is at the message transport network interface and the other is in the message content and format. These standards are needed to attain the desired ubiquity for CMS. The vast variety of different users in different countries speaking different languages and working in different disciplines using different user interfaces will need to adhere to some minimum format and content standards if they are to exchange messages via some common transport network. They will also have to adhere to standards for the content of their messages to allow receiving computers to correctly interpret the content. Certain fields of each message will have to adhere to format standards. Development of these standards will require consideration of both user and system environment issues. Page 12 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT Among the payoffs that users will expect from standards will be regular ways to represent the full alphabets of different languages and disciplines to enable communication in every possible language. In return for conformance to CMS standards, the users should see an expanded range of communication and greater ease of access to, and use of, CMS facilities. F. LONG TERM USER EXPECTATIONS Computer Message Services that handle textual information represent an early manifestation of the convergence of computers and communications. In the long term, with continued technology advances and continued convergence, CMS should expand to handle voice, graphics, facsimile, and other forms of data in CMS facilities. The expectation is that users, with appropriate equipment, should be able to send and receive messages that "contain" all these forms of information. G. USER ECONOMICS - COSTS VS BENEFITS In the user environment, economic issues are dominated by questions of "cost/benefit." Will the customer derive enough benefit from use to justify purchase of the service. In the long term, we must ask whether CMS can be and will be viable as a competitive service. Since Computer Message Services are not yet well defined, it is difficult to answer this question in advance. Instead we must ask what kinds and levels of end-user service will promise to make CMS viable in the international marketplace. CMS appears to promise the following kinds of communication benefits: Connectivity among individuals and agencies with access to CMS. Those who have access will be able communicate through the CMS facilities with others who have access. As the number of connected individuals and agencies grows, the value of being connected will grow. Evidence of this value can be seen in the ubiquitous TELEX/TWX service, which derives great benefit from having a large number of terminal nodes installed in every nation. This same kind of value is evidenced where voice telephone service is widely available. The same can be said for regular postal services which serve large populations. The key source of value lies in the range of easily addressable potential communicants. Page 13 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT Mobility of information under individual or agency control. With access to CMS, users will be able to move information from their computers to other CMS users, and thus be able mobilize their information and deploy it as may be needed. Potential ease of controllable movement will provide important benefits. This raises the issue of security and privacy. Information value can be diminished by undesired disclosure or unauthorized access, as by enemies, or competitors, or even innocent misdelivery. One key to user values derived from mobility lies in the ability of an individual or agency user to control the delivery of information sent through the CMS facilities. Page 14 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT VI. CMS SYSTEM ENVIRONMENT DISCUSSION System Environment Models are generally concerned with architecture and with the protocols and interface standards for interactions among components and modules, initially cast in terms of services to be delivered by the system. Workshop discussions identified the following list of system services: TRANSFER, MEDIATION, STORAGE, RELIABILITY, SECURITY and PRIVACY. The desired CMS System Architecture was seen to combine these service facilities to provide an interface to user environment facilities for sending and receiving computer messages. A. ARCHITECTURE Workshop presentations identified the following kinds of components: computers, terminals, communication links, communication processes, interprocess interfaces, message transport mechanisms, file systems, user interfaces, computer operating systems, message delivery mechanisms, address interpreters, routing mechanisms, security controls, encryption mechanisms, accounting systems, directory data bases, mailboxes, computer networks, record carriers, etc. Now the question becomes one of deciding how to organize the collection of such things into a workable system to provide international computer message services. Discussions led toward a conclusion that the architecture for CMS should be open to accommodate the widest possible kinds of user facilities, and thus provide a solid technical basis for CMS to become ubiquitous. The system architecture should also accommodate existing message system interconnection, and facilitate utilization of existing message system facilities. Except for identification of probable components and general notions of architectural constraints, the workshop did not develop specific architectures or models. A number of concepts were presented calling for modular separation among send/receive processes, actual transmission processes, and user processes. Such an architecture demands distributed processing and decentralized control, which would appear to be appropriate for an international system based on distributed ownership and operation of facilities. Further development of architectures and models was left to the SYSTEMS ENVIRONMENT Subgroup. Page 15 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT B. PROTOCOLS Protocols are the syntactic and semantic procedures used between components of a distributed system to provide linkages and interfaces among the modules which collectively perform the service functions to be delivered by an application system, in this case CMS. Since CMS protocols would control communications across both vendor and organizational boundaries, agreements on standards for protocols to support CMS are essential if service ubiquity is to be realized. Various protocols were discussed in the workshop, ranging from research proposals for inter-network computer message services, to the developing CCITT standards and protocols for TELETEX services. ARPANET message systems, based on file transfer protocols were discussed, and it became clear that currently operational ARPANET mechanisms do not represent the desired protocols for an international CMS system, though ARPANET experience has been quite valuable. Another developmental protocol was described which uses the dial telephone network, (and might utilize any available communication channel) to establish a message transfer facility between computers, but it it was presented as an interesting expedient solution for near term implementation. It was recognized that messaging protocols utilize basic data transfer mechanisms such as generalized inter-process communication, file transfer, and virtual terminal services. The issues in these areas have been under study by IFIP WG 6.1 for a number of years and work on developing international standards is in progress in ISO (Open Systems Architecture) and CCITT (Layered Models Of Public Data Network Service Applications) and it is presumed that these would provide the underlying mechanisms for message transfer etc. The development of message architectures and models by the Systems Environment sub-group should lead to the identification of specific protocols for message services that need to be standardized by either CCITT or ISO or both. Page 16 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT C. SERVICES 1. TRANSFER Transfer services would be provided by implementation of protocols in distributed intercomputer processes to achieve delivery of "posted" messages through some communication link between sending and receiving computers. Transfer processes might operate as directly invoked user processes, or as background processes not requiring direct involvement of either the sender, the receiver, or their terminals. Transfer might use simple links for proximate computers, or complex internet links for distant computers. Distance in this context must be measured in terms of the available network paths rather than geographic distance. Although discussion showed no inherent reason why transfer services should not also allow delivery of messages to relatively low capability terminals such as the new TELETEX or older TELEX terminals, delivery to computers was considered essential to any CMS. Likewise, origination and sending from low capability terminals could not be disallowed. Delivery to a computer would typically mean that the receiving computer creates a file copy of the message for retrieval by the user upon request. In concept, the CMS delivery process would deliver to a "mailbox" in a computer and the user would poll the computer mailbox for new arrivals. Discussion brought out problems with routing of transfers across multiple networks, through gateways or other facilities. One of the basic difficulties here had to do with naming confusion. Gateways have taken on specific technical meaning in the development of internet protocols, and IFIP WG 6.5 would like to avoid conflicting use of terms. During discussion, the term "message transfer station" was coined for use to identify the mechanism for transferring messages across internet boundaries. Transfer services were seen to require end-to-end control of transfers to assure proper delivery to the intended addressees. Page 17 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT Addressing structures of several messaging systems were described. The discussion noted a need to clearly differentiate between the notions of name, address and route. In addition, researchers were reaching the conclusion that it is architecturally important to differentiate between naming-addressing aspects of the transfer service and the structure and contents of a message. The notion was introduced of a property list (somewhat analogous to a postal envelope) which is to be available to both the transfer facilities and to the message recipient and is guaranteed to be accurate. This property list should be used by USER ENVIRONMENT processes for REPLY generation and other functions that critically depend upon accurate information about addresses, postmarks, identification numbers,, etc. This distinction between "envelopes" and "contents" is essential for generalized message services where the contents may not contain any addressing data. eg. pictures, programs etc. 2. MEDIATION Mediation services are needed to allow interfacing among diverse and often incompatible terminals, computers, networks, and services. Voice telephone and TWX/TELEX connections leave mediation to be done by the connecting devices when they are not naturally compatible. Some networks such as TYMNET, TELENET, DATAPAC, TRANSPAC, and PSS (among others) provide mediation among diverse terminals and HOST computers through the net, but they do not yet provide universal mediation for all possible computers and terminals. And these networks do not currently provide general mediation services between incompatible computers for file transfers or message transfers. The ARPANET provides, in a research environment, a greater array of mediation services than most other networks, and thus ARPANET experience has been important in developing the concepts. However, as with any artifact of research, the ARPANET does not generally provide a good example of how to implement any specific service function. Instead, problems discovered from ARPANET experience will more often point to better ways to implement in successor systems. Additional complexities are seen for mediation services in trying to mediate between foreign language alphabets, and foreign languages in general. A companion problem of similar complexity will arise from the need to accommodate older existing message facilities and terminals. Page 18 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT 3. STORAGE Storage services in the CMS systems would serve several functions. One is to hold messages enroute to their destinations, either while being transmitted, or while backed up waiting for a channel to open. Other purposes might be to provide archival storage for authentication, or for long term storage. In certain cases, encryption keys might need to be held in CMS systems storage for an encryption system to serve a large public. Other uses would include public directories such as the name/address data bases. 4. RELIABILITY Reliability services are needed for rather obvious reasons. If users are to become dependent on CMS services, then the service must be reliable in terms of guaranteed delivery, continuous operation, ability to accommodate to address changes, adapt to alternate routing when channels are blocked, etc. And when all else fails, the CMS systems must notify the sender of failures, or at least provide such notice when requested by the sender. To provide these functional services, CMS will need to use redundant communication channels, adaptive routing schemes, error control mechanisms, end-to-end transfer control, and provide records of deliveries. Some form of sender authentication should be included, but workshop discussion led to the notion that CMS should probably use "out of channel" communications to provide positive authentication at acceptable costs. There is some evidence that stringent requirements for absolute authentication of senders might involve only a small fraction of CMS messages. 5. SECURITY & PRIVACY There were Workshop presentations and discussions on encryption mechanisms aimed at security and privacy, including transmission of "signatures" and other forms of authentication. By and large, encryption schemes were characterized as relatively expensive on the one hand, and of somewhat uncertain future guarantee against cracking on the other. Most schemes depend on computational difficulty for their protection against code breaking, and there has been a history of mathematical breakthroughs to render some schemes relatively insecure. It is impossible to know when in the future someone will achieve another breakthrough to short cut the computational complexity of other schemes. Page 19 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT Another problem with encryption for security and privacy stems from the problem of key establishment and distribution. Various schemes for public keys and other means for handling keys were discussed, but no conclusions could be drawn for CMS system designs. This is not to say that encryption is useless or pointless. Instead, it seemed clear that use of encryption should be limited in CMS to those situations where its high cost can be justified. None-the-less, security and privacy remain serious problems. Some national governments insist on the privilege of inspection for electronic message transmission, and there exist legal requirements to disclose to some governments, any keys used to encrypt electronically transmitted information. D. STANDARDS System environment standards are needed to achieve interoperability among the diverse distributed processing modules that will make up an international message system. Standards are needed in the areas of transfer facility interfaces, naming & addressing, dates, content structures and formats. Recommendation X.121 proposes an international numbering scheme for public data networks. Interoperability here means that the sent and received messages should be computer processable by both parties and by the transfer facilities between them. At present, most of the standards effort is focused on text messages, but it is clear that future developments will soon need efforts focused on mixed media including voice, facsimile and graphics. E. ECONOMICS - COSTS VS PRICES SYSTEM ENVIRONMENT economics are basically a profitability question. Can the required kinds and levels of service be provided at acceptable prices using system facilities that will yield acceptable profits (or losses)? Cost factors include terminals, computers and communications, plus security, reliability, ubiquity, etc. Page 20 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT VII. MS POLICY ENVIRONMENT DISCUSSION A. REGULATION The regulatory policy environment appears to be quite complex due to wide variations in prevailing attitudes and practices in different countries. The USA allows interconnection to public telecommunication networks while other countries do not. In some countries, communication utilities are combined with the regular postal service in the form of government owned and operated PTTs. Even the meaning of the word "regulation" is different from one country to the next. And, in the USA there is movement to rewrite the entire basis of regulation with legislative action to facilitate competition. This means that regulatory attitudes and actions regarding CMS will be hard to predict. A great deal of education may need to be done, with a great deal of negotiation to follow. Emergence of CMS at this time will add to the difficulty of sorting out the issues, and it will be sensible to make CMS part of the solution rather than part of the problem. B. TARIFFS Tariffs serve more than one role. They provide the means for dividing revenues among international partners, and provide means for controlling the flow of information across international borders. They also provide for subsidies in various forms, and serve to control the kinds and levels of service to be offered in any given state or country. Several models for CMS tariff patterns are available, but it is uncertain which might apply. Postal Service models involve third party delivery agencies of the form envisioned for some CMS services, while the TELEX and Voice Telephone models involve international communication channels as required by CMS. Even Broadcast models have some applicability. The task presented to the 6.5 POLICY ENVIRONMENT Subgroup is to look for some ways to draw a blend of applicable models to form a basis for CMS tariffs. Among the things to be accomplished is analysis of the available models, and comparison of each with the evolving conceptual models of CMS. Page 21 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT C. USE OF EXISTING FACILITIES It is imperative that existing investments in international message communication facilities be protected, while it is equally imperative that CMS facilities be developed, installed, and brought into operation. Thus ways must be found to utilize existing facilities in the evolution of CMS. Policies should be explored for collaboration with the operators of existing facilities. D. SPECIFICATION OF FUTURE FACILITIES Without concurrence among policy bodies among the affected nations and states, it is impossible to foresee full network interconnection of CMS facilities. Somehow a scenario for full interconnection needs to be developed. E. ECONOMICS - TRADE AND COMMERCE With the advent of computer networks, and now the promises of CMS, the information processing industry is beginning to transform itself into a modern mass production, mass distribution, mass consumption industry. Without benefit of computer networks, production of information services has been confined to co-location with local markets. With computer networks,, and with CMS, information service producers may gain access to mass markets, and consumers may gain access to the full range of produced services. In short, CMS is part of a potentially burgeoning international industry, with information serving as the basic commodity, computation serving as the basic engine of production, and communication serving as the basic means of transport. All the benefits of trade can accrue to participants in this new industry structure, if the trading partners can work out a policy framework to facilitate trade. Resolution of the issues will not be easy. Information does not behave like other commodities. For example, use does not necessarily diminish supply, and the value of information may be either enhanced or diminished by disclosure, depending on each specific case. Protection of both private and national interests in information will be important, but it will not be easy to determine how to protect those interests. Page 22 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT As with other cases of industrial transformation from cottage industry to modern industry structures, there will be considerable disruption of the old organizations of capital and labor. This was noted in the discussion about protection of the existing message system facilities and investments. The same concerns must be applied to problems of labor displacement. Policies for accommodation to change will be needed in every sector that will be impacted by CMS, and it is difficult to imagine any sector that might be left out. Page 23 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT APPENDICES The appendices include the list of attendees, tentative subgroup participants, and addresses of IFIP WG 6.5 Chair people. APPENDIX A: ATTENDEES Belgium: Andre Danthine; Canada: George Barker, Gregor Bochmann, Ian Cunningham, Gwen Edwards, Bill McLean, Penny Napke, Dave Twyver, Ronald Uhlig; France: Bernard Gagey, Michel Gien, Jean-Louis Grange', Najah Naffah, Louis Pouzin, Hubert Zimmerman; Germany: H. Fergen; Switzerland: Peter Schicker; UK: Mike Burgess, Donald Davies, Bill Medcraft, Roger Pye, Terence Westgate; USA: Vint Cerf, Lennie Copeland, Ira Cotton, David Crocker, John Day, Debbie Deutsch, David Farber, Elizabeth Feinler, Ken Harrenstein, Ron Kunzelman, Robert Metcalfe, John Peters, John Pickens, Jon Postel, Mike Schroeder, Einar Stefferud, Philip Walker, Shirley Watkins, Jim White. Page 24 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT APPENDIX B: SUBGROUP PARTICIPANTS USER ENVIRONMENT SUBGROUP: Chair: Donald Davies (UK) & Elizabeth Feinler (USA) Tentative Participants: Belgium: Andre Danthine; Canada: Ian Cunningham; France: Bernard Gagey, Jean-Louis Grange'; Germany: H. Fergen; Switzerland: Peter Schicker; UK: Donald Davies, Roger Pye, Terence Westgate; USA: Vint Cerf,Ira Cotton, David Crocker, Debbie Deutsch, David Farber, Elizabeth Feinler, Robert Metcalfe, Mike Schroeder, Einar Stefferud, Shirley Watkins. SYSTEM ENVIRONMENT SUBGROUP: Chair: Najah Naffah (France) & Ian Cunningham (Canada) Tentative Participants: Belgium: Andre Danthine; Canada: Gregor Bochmann, Ian Cunningham, Gwen Edwards, France: Michel Gien, Najah Naffah, Hubert Zimmerman; Germany: H. Fergen; UK: Bill Medcraft; USA: Vint Cerf, Ira Cotton, David Crocker, Debbie Deutsch, David Farber, Elizabeth Feinler, Ken Harrenstein, Robert Metcalfe, John Pickens, Mike Schroeder, Shirley Watkins, Jim White. POLICY ENVIRONMENT SUBGROUP: Chair: Gwen Edwards (Canada) & Roger Pye (UK) Tentative Participants:, Canada: Gwen Edwards, Ronald Uhlig; France: Louis Pouzin; UK: Mike Burgess, Roger Pye, Terence Westgate; USA: David Farber, Philip Walker. Page 25 30-Sep-79 MARCH 1979 IFIP WORKING GROUP 6.5 WORKSHOP REPORT APPENDIX C: ADDRESSES OF IFIP WG 6.5 CHAIR PEOPLE Cunningham, Ian (Co-Chair: System Environment Subgroup) Bell Northern Research PO Box 3511, Station C Ottawa, Ontario, Canada K1Y 4H7 Davies, Donald (Co-Chair: User Environment Subgroup) National Physical Laboratory Teddington, Middlesex, UK Edwards, Gwen (Co-Chair: Policy Environment Subgroup) Bell Canada 25 Eddy Street Hull, Quebec, Canada Farber, David (Chair: Workshop Program) University of Delaware Dept. of Electrical Engineering Newark, Delaware, USA 19711 Feinler, Elizabeth (Co-Chair: User Environment Subgroup) SRI International Network Information Center, J2021 Menlo Park, CA, USA 94025 Naffah, Najah (Co-Chair: System Environment Subgroup) IRIA 78150 Rocquencourt France Pouzin, Louis (Chair: IFIP TC 6) IRIA 78150 Rocquencourt France Pye, Roger (Co-Chair: Policy Environment Subgroup) Communications Studies and Planning Ltd. Circus House 21 Great Pitchfield London, UK W1P 7FD. Uhlig, Ronald (Chair: IFIP WG 6.5) Bell-Northern Research PO Box 3511, Station C Ottawa, Ontario, Canada K1Y 4H7 Page 26 30-Sep-79 IFIP WORKING GROUP 6.5 REPORT OF THE MARCH 1979 MONTREAL WORKSHOP 30 Sep 79 -------