INTERNATIONAL TELECOMMUNICATION UNION OSI & Internet Seminar Tuusula, Finland 21 June 1994 OSI & Internet: Competition or Cooperation? Presentation by Dr Pekka Tarjanne Secretary-General, International Telecommunication Union "OSI & Internet; the makings for a perfect marriage?" This conference has a very auspicious title - or let me say it proposes an enticing speculation. Although I am sure that most of you are well acquainted with both, allow me to briefly introduce the prospective bride and groom. The Internet - with an upper-case "I" - is a collection of thousands of networks joined by physical communication links and a common set of technical protocols which make it possible for users of any one of the networks to communicate with or use the services located on any of the other networks. These protocols are referred to as TCP/IP, the TCP/IP protocol suite, or more correctly now: the Internet Protocol Suite (IPS). The Internet started with networks in the United States, but now includes networks in many parts of the world. Some of the networks linked in the Internet have supported multiple protocols for a long time. The "Internet Standards were once limited to those protocols composing what has been commonly known as the "TCP/IP protocol suite". However, the Internet has been evolving towards the support of multiple protocol suites, especially the Open Systems Interconnection (OSI) suite" . The TUBA effort has already introduced CLNP, co-existing with IP, in numerous wide area networks which are part of the Internet . Products which comply with the IPS are widely available, and TCP/IP support is commonly included with the operating system software for many computer systems. Consequently IPS is used not only on the Internet, but also in thousands of private networks. Open Systems Interconnection (OSI) stems from the realization in the late 1970's that there was no master plan or structure for determining what standards were needed for data communications and distributed information systems. It was generally recognized that vendor-specific implementations of data communication protocols led to isolated domains of information, which were very difficult and expensive to bridge, and that International Standards provided a basis for open connection and interoperation of equipment and communication services. In 1977 the International Organization for Standardization (ISO) Technical Committee 97 established a new subcommittee, then designated as SC 16, entitled "Open Systems Interconnection." SC 16 developed an architecture to provide a framework for the definition, development, and validation of standards in the new generation of distributed information systems. The ITU's CCITT also recognized the importance of establishing a sound architecture and appointed a Special Rapporteur to study the problem and to work in close liaison with ISO . This work resulted in the well-known seven layer Reference Model which is both an ISO International Standard, ISO 7498, and CCITT Recommendation X.200. Standards defining services at all layers of the Reference Model have since been produced. Many manufacturers and telecommunication service providers offer products and services complying with OSI standards, and many governments have mandated compliance with OSI standards for data processing equipment procurements. There is a widely held view that the Internet Protocol Suite now realizes many of the OSI objectives, particularly in making it routinely possible to interoperate off-the-shelf systems from different sources. Therefore the marriage of OSI and the Internet Protocol Suite is attractive. If they are not already married, at least they have been getting together, insofar as there are examples of use of OSI standards in the Internet (for example, X.500). Users, particularly in government and in large corporations doing business internationally, have a preference for recognized international standards and a very strong preference for convergence of standards, so as to facilitate interoperability and avoid unnecessary costs. Standards Process Many distinguished experts will be addressing you today, so I do not intend to expound on the use of particular standards from the OSI suite in the Internet. Still less would I speculate on the place for CLNP in IP: the next generation (known as IPng). So leaving technical matters for others, let's turn to the standards process. The Internet community has given rise to a successful standards making process that has served it well. The process is characterized by: * practical testing of proposed standards: at least two interoperating implementations are generally required before adoption of a protocol specification as an Internet Standard; * openness: free public access to documentation, including working documents; * extensive use of electronic communications for speedy, low-cost inclusion of all interested parties. There is now a conscious move towards increased formality without sacrificing these advantages. The legal basis of the efforts is the Internet Society which calls itself "the international organization for global cooperation and coordination for the Internet and its internetworking technologies and applications." Standards decisions are made, without formal voting to assure consensus within the IETF, by the Internet Engineering Steering Group (IESG). The IESG is appointed and overseen by the Internet Architecture Board (IAB), whose members (not more than 15) are nominated by the IAB itself, by the President of the Internet Society, or its Board of Trustees, or exceptionally by petition from the membership of the Internet Society1, . IETF standards development has its roots in an informal consensus process. This has been a strength through its emphasis on working code prior to standardization and its ability to rapidly adapt to the changing community requirements. However, it has resulted in a perceived lack of formal mechanisms for achieving balance between various interested parties . The recent final report of the U.S. inter-agency Federal Internetworking Requirements Panel (FIRP) points out that "There are some areas, such as standards for addressing schemes and administration over allocation of addresses, for which it is extremely important to have a single standard and process to provide worldwide connectivity. The international community and other governments will only accept formally recognized international and national organizations dealing with such matters." There are some types of issues for which a formal international legal framework may be even more important, for example: operating and service reliability questions, tariffs and accounting principles. The ITU-T study groups deal not only with technical questions, but also with these questions. Some matters simply need to have the legal foundation that formal international instruments, the equivalent of treaties, confer. The ITU is an intergovernmental organization with 184 member countries. Telecommunication service providers, manufacturers, scientific and other international or regional organizations also participate in ITU activities such as standardization. At present only the Member countries have the right to vote, but the standardization process is driven by the private companies which contribute more than 85% of the input documents. Participation in ITU is institutional. Study Group members are "delegates" sent by their Administrations - the ITU term for the Member countries, or "participants" from other organizations admitted to the standardization activities of ITU-T. This contrasts with the individual role of persons who participate in Internet standardization activities. The ITU model assures representation of the interests of economically - and politically - important entities. There is also a continuity implicit in institutional participation - although Vint Cerf and Jon Postel are certainly institutions in their own right! ITU's history goes back to May 1865 when the first International Telegraph Convention was signed by the 20 participating countries and the International Telegraph Union was established. The International Telephone Consultative Committee (CCIF) set up in 1924, the International Telegraph Consultative Committee (CCIT) set up in 1925, and the International Radio Consultative Committee (CCIR) set up in 1927, were established to coordinate the technical studies, tests and measurements being carried out in the various fields of telecommunications and to draw up international standards. In 1932 the Telegraph and Radiotelegraph conventions were combined, and in 1934 the ITU name was changed to become the International Telecommunication Union in order to reaffirm the full scope of its responsibilities, i.e. all forms of communication, by wire, radio, optical systems or other electromagnetic systems. In 1956, the CCIT and the CCIF were amalgamated to give rise to the International Telephone and Telegraph Consultative Committee (CCITT). The standardization work in the former CCITT, and now in ITU-T, is done in Study Groups. Each Study Group deals with a specific area of telecommunications such as: transmission, switching, data networks and open system communications. The CCITT carried out its work in four year cycles culminating in a Plenary Assembly where Recommendations developed or modified by Study Groups would be approved - traditionally unanimously. Since 1988 this time-consuming and costly method is now more frequently replaced by a "fast-track" procedure in which draft standards, having been agreed upon unanimously in a Study Group meeting by Administrations, are then given final approval by correspondence if 80% of the replies received are positive. The standards are then immediately published in separate editions, a departure from the former CCITT colored Books which followed the four year cycle. This accelerated procedure greatly speeds up the standards production cycle . By 1989 the ITU's structure, working methods, and procedural rules which had evolved over the years no longer fit the rapidly changing telecommunications environment. The 1989 ITU Plenipotentiary Conference mandated a High Level Committee to thoroughly study the reforms needed. The restructuring of the International Telecommunication Union recommended by the High Level Committee was officially decided upon by a special Additional ITU Plenipotentiary Conference (APP-92) held in December 1992. The new structure took effect on 1 March 1993 and includes a single Telecommunication Standardization Sector which consolidates the activities of the CCITT and the standards-setting activities of it's former radio counterpart, the CCIR. The spirit of reform continued with the first World Telecommunication Standardization Conference (WTSC-93), held in March 1993 in Helsinki, taking the place of what would have been the Xth CCITT Plenary Assembly. Among the items dealt with by the WTSC was the creation of the Telecommunication Standardization Advisory Group (TSAG). This sort of "standards making management group" reflects the importance of the private sector in ITU standardization. TSAG provides guidance for a market-oriented, instead of technology-oriented, approach to standardization, with Study Group programs identifying priorities and delivery dates for standards. ITU has also taken steps to increase the openness of its processes and to make the processes more efficient with extensive use of Electronic Document Handling (EDH) and public electronic access to documents. The Global or International Information Infrastructure (triple-I) There has been much discussion recently about the Global or International Information Infrastructure (the triple-I), and there are differing views of what the triple-I might be: * At one level it is seen as a high performance computer network which will facilitate high speed data access and retrieval. In this model, the Internet is sometimes seen as the precursor for the International Information Infrastructure. * An alternative definition is a multimedia network for which the primary use will be conveying video datastreams in conjunction with data, text and voice. According to this vision, many of the potential applications will be in the entertainment, education and health care as well as the business market. * A third definition is as a medium for interactive television, in which it is the intelligent television set rather than the home computer or the videophone which is the main communication channel. The battle for the television set-top is likely to be every bit as dynamic as the battle for the desktop over the coming years as different technological solutions are proposed to help consumers cope with a rich diet of new television channels, video-on-demand, home shopping and other services. Teenagers playing video games could be using the network alongside multinational corporations holding video- conferences. It is not surprising that these three visions of what the International Information Infrastructure should become come from different parts of the information industry as it currently exists: the computer industry, the telecommunications industry and the broadcasting industry. But the beauty of modern technology is that a single network of networks can, in theory, accommodate each of these different applications. This single network will be digital. The process of digitization began in the computer industry, it is already well-advanced in the telecommunication industry and it is now spreading to the broadcasting sector. As the three sectors converge, it will become increasingly difficult, and unnecessary, to distinguish between the different parts of the bit business. Information should, in theory, be able to flow from any source to any destination providing the network is digital, and providing the addressing scheme is universal. Some say that the triple-I could be an outgrowth of the Internet; others believe that it will be an evolution of the conventional public network. These alternatives are not mutually exclusive, since the Internet is built on top of the public structure. The common network infrastructure which can accommodate the different applications - including computer internetworking - will have to be built according to underlying standards which meet broad needs. The infrastructure standards will have to reflect the requirements of these diverse applications. Implicitly, the standards process for the high performance computer network must mesh with the overall international standards process. There is nothing new or surprising to this. The Internet has always run over publicly supplied infrastructure which in turn conformed to standards from telecommunication sources - proprietary Bell System standards during its early history; in recent years, regional and international standards. The recent RFCs for IP over ATM and for PPP over Sonet/SDH are examples of how such standards are incorporated in the Internet. There is reason to question whether even the high performance computer network part of the triple-I will grow from the Internet - or will a parallel structure be built up? A number of new internetworking service offerings based on public network infrastructure have been introduced or announced recently. Right now a great deal of information - mostly free information - is on the Internet. And so are millions of attached systems (2.3 M) and many millions of users. But there are formidable challenges which include dealing with business issues on an international scale. * Operational matters. Traditional telecom operators guarantee a very high level of reliability - which is a must for business users, and which private users also take for granted. * Interests of established large operators. These are the investors who own the network plumbing: the vast fiber network which has grown up over the last decade. * Geographic reach. A truly global system - like the telephone network - must provide service in all countries of the world. * Political issues. The Internet Society is non-political, which may be an advantage for technical matters, but some questions related to telecoms development and relations among operators are inherently political - or at least have political dimensions. * Regulatory aspects. Regulatory regimes differ throughout the world for enhanced or value-added service offerings, which is what internetworking services might be considered under some regimes. * Tariffs, accounting and international settlements require agreed- upon equitable bases for identification of costs and value of services rendered. These are issues for which the international telecommunications community turns to ITU as the formally constituted international body. Competition or Cooperation in Standards Competition or Cooperation? There will be both. Competition is healthy, and keeps everybody on their toes. Just as in competitive telecommunications environments, competition in standardization can lead to the best technical solution and to better performance. Organizations know that untimely, poor results will cause the customer to go elsewhere - because there is an elsewhere. Exhortations to competition in standardization are not needed. We'll have competition regardless of what the Secretary-General of ITU might say. But I am also a strong advocate of sincere and effective cooperation among standards organizations. Cooperation is in the best interests of industry, users, and the standards process itself. Those of you who participate in the work of SC 6 know of the long- standing cooperative relationship among ISO, IEC and CCITT, now between JTC1 and ITU, with a rational division of work - and where SC 6 and SC 21 have contributed so much. The key point in standards cooperation is to avoid duplication of work, waste of resources - standardization experts are the scarcest resource - and especially conflicting standards. JTC1 and ITU-T have different mandates, which makes coordination easier. The regional telecommunication standardization organizations (RSOs) such as ETSI in Europe, T1 in North America, and TTC in Japan deal with genuine regional topics, but also with large areas which overlap the work of ITU-T. Basic coordination and cooperation mechanisms were agreed upon among ITU and the RSOs at the Interregional Telecommunications Standards Conference (ITSC) in February 1990 in Fredericksburg, Virginia. Subsequent ITSC meetings have resulted in increased collaboration on mutually agreed upon areas of high, continuing common interest and agreement on principles such as openness and "adoption of work." The ITSC process has now given birth to a new structure known as the Global Standards Collaboration (GSC) group which held its first meeting in March 1994 in Melbourne. Cooperation with industry groups, such as the ATM Forum, also promises to meet industry and user needs (e.g., for new AALs) rapidly, while providing input to the ITU-T work on B-ISDN. There are calls for cooperation on internetworking matters, with particular attention focused on the network and transport layers. The Federal Internetworking Requirements Panel report which I cited earlier suggests "convergence of both IPS and OSI to [a] new internetworking layer" and goes on to recommend "that the IETF and JTC1 SC 6 jointly establish convergence workshops that take advantage of the best characteristics of both organizations"7 But I promised to leave technical questions for the experts who will be speaking after me. So I want to emphasize the importance of cooperation in domains that may fall outside SC 6's preoccupations. On the one hand, there is the infrastructure itself where Internet service providers are users - and where I would like to encourage active participation by the Internet community in the ITU groups developing standards for technologies like SDH and ATM. On the other hand, there are questions concerned with tariffs, accounting and settlements, and with service levels and other operational matters. These are issues that have to be addressed to respond to real concerns of users and service providers. Some of these questions require arriving at legally binding arrangements, generally at an international level. Satisfactory resolution of these questions is needed for international commercial internetworking. There are those who will say that for one reason or another these issues are not so important in the Internet environment. For example, that accounting - especially in a connectionless context - is too costly to be viable, and that the Internet has gotten by until now with essentially flat rate charging schemes. It may be that the basis for settlements may be more economically calculated with aggregated traffic statistics, than with per call accounting. But the question of appropriate charging paradigms is not an engineering decision. It is one for users and service providers to resolve in the marketplace. Arriving at a methodology for identifying cost elements in complex networks in which multiple operators participate is a difficult task. Much work has been done under ITU auspices in defining accounting principles in this area. For example, ITU-T Study Group 3 has recently revised Recommendation D.36 which provides the general accounting principles applicable to message handling services. Recommendations such as D.36 provide guidelines, but the specific details of each particular accounting arrangement are the subject of bilateral agreements and may vary from the Recommendation, leaving latitude for operators to make commercial arrangements which are appropriate to their circumstances. Conclusion There is an Area in the IETF which promotes extension of the Internet in developing countries. Cooperation in this domain with ITU's Telecommunication Development Bureau (BDT) could be fruitful. The March, 1994 World Telecommunication Development Conference established a work plan for BDT which includes a programme for development of telematics and computer networks. There already is standards-related cooperation, especially in the form of the Internet community adopting International Standards. There are individuals who have made contributions to the OSI effort in International Standards bodies who participate in IETF work. ISO, IEC, and ITU are users of the Internet; and ITU documentation is available electronically on the Internet. But it would be desirable to have increased participation by the Internet community in the International Standards processes, so that Internet needs and preoccupations are reflected in the work of ITU and JTC1 bodies. This would be beneficial to both Internet and the international standards organizations. It is also possible to envisage a mode of cooperation similar to that which works well with the regional standards organizations (RSOs), in which standards for certain areas would be developed by the IETF and would reach JTC1 or ITU-T as proposals. Just as is the case for consolidated RSO standards proposals, they would be open for discussion and, if necessary, modification and alteration. Perhaps the metaphor of a marriage is not quite right for the relationship between the International Standards organizations and the Internet, but certainly there is potential for constructive partnership. _____________________________ Huitema, C. and Gross, P. "Internet Standards Process -- Revision 2," RFC 1602, March 1994 Ford, P.S., Rekhter, Y., Knopper, M., Colella, R., "TUBA: CLNP as IPng," ConneXions, May 1994. Folts, H., "A Tutorial On the Open Systems Interconnection Reference Model" in IEEE Computer Society Tutorial "Computer Communications: Architectures, Protocols, and Standards," 1987 Fleischman, E. "A User's View of IPng," ConneXions, September 1993 Cerf, V., "Charter of the Internet Architecture Board (IAB)," RFC 1358, August 1992 Leiner, B. "Internet Draft IAB Coordinated Response to Draft FIRP Report" February 1994 NIST, final report of the Federal Internetworking Requirements Panel (FIRP), May 1994 ITU Convention, Geneva, 1992, art. 14 No. 193 Irmer, T., "Shaping Future Telecommunications: The Challenge of Global Standardization," IEEE Communications Magazine, January 1994 ITU Council, Document C94/60, Annex 2, "Buenos Aires Action Plan for the Global Development of Telecommunications," April 1994 OSI & Internet: Competition or Cooperation - 7 -