Internetworking: Critical Information Infrastructure Keynote Address The Network Services Conference 1993 Warsaw, Poland 12-14 October 1993 Anthony M. Rutkowski Vice-President, Internet Society Director, Sprint International On the way to this beautiful city in the homeland of my grandparents and countless generations of ancestors, I had the opportunity to visit a former institutional home in Geneva - the International Telecommunication Union. There, the Secretary-General, other Union officials, and staff from nearby international organizations invited me to make a presentation and discuss with them "the internetworking revolution." Many of these organizations are now heavy users of the Internet. They all wanted to know: What is happening? Why is it happening? It's a common phenomenon these days. Last week, Wall Street Journal Television - which is doing a special program on the Internet - visited Internet Society headquarters asking the same questions. The week prior to that, the World Bank held its first Internetworking seminar at its Washington headquarters and announced that the case was so compelling for its implementation and use, that numerous new internal and external initiatives were being rolled out. Even the annual almanac of the telecom- munication world - TeleGeography - is inserting a special section on internetworking for the first time in the 1993 edition. Someone recently did a string search for articles in major USA public newspapers and magazines and come up with the remarkable statistic that there were more than 170 articles on Internet during this past summer. My standard initial reply to queries about this subject is simple but dramatic - Internetworking is one of the most revolutionary technologies of the 20th century; Internetworking is perhaps THE most revolutionary human communications medium that has yet emerged. Why is this so? What is it so remarkable about internetworking? Is this exaggerated? I think not. And, the answer is not just that it's an important technology and major new business sector, but rather it's a critical enabling infrastructure for institutions, for professions, for people, for countries, for creative genius itself. In many respects, it is the enabling side of this phenomenon that is more important than the technology. Internetworking is a destroyer of time and of space. It removes institutional walls. It is a changer of paradigms. Indeed, this spirit of individual enablement is reflected in the marvelous titles of some of the presentations at this conference: Sowing the Networking Seed, User Network Interface to Everything, Networks as Tools for Hunting Historical Treasures, Networked Collaborative Environments - to name but a few. What I hope to share with you today is a an overview of current global developments to emphasize the importance of this conference in furthering the internetworking paradigm, and the significance of your individual and collective efforts in bringing it about. It shouldn't escape our notice that Warsaw is certainly an appropriate venue for this focus, for it was only a few years ago that the actions of other motivated people meeting in this city and country brought about another kind of paradigm shift benefiting both individual freedom and critical enabling infrastructure! ***** Internetworking is its own revolution In it's broadest manifestation, the Internet is a global mesh of information, information processes, and people. It is flat information space where any organization, person or computer can almost instantly discover, receive from, or transmit to - any other organizations or persons or computers with scalable high performance and low cost. Nearly twenty million people, one thousand processes and two million computers - now exist in one vast worldwide information space. The accompanying first set of slides are intended to provide a flavour of the measurable dynamics and trends of the internetworking environment today. The real significance of all this lies not so much in the technology and the growth of internets, but in what's happening "on top" of this mesh - that is, what institutions and people around the world are doing with these capabilities. Nonetheless, because "connectivity is its own reward," these metrics do provide a measure of the scale and pace of the activity that is captured in the slogan The Internet is its own revolution At the outset, it is important to emphasize that the revolution taking place is not solely limited to a specific technology or implementation. Because The Internet is literally the connectivity among many diverse networks, it is all of them collectively that provide the aggregate benefit. Unfortunately the statistics don't really account for the additional growth of the rest of the broader Internet "Matrix" which includes Bitnet, UUCP networks, Fidonets or the gateways to most major public and private messaging systems in the world. Collectively they now provide access among 137 countries and territories. IP Internets. We shouldn't overlook that even the implementation of presently unconnected enterprise internets results in major scales of economy, and provides those independent internets with "plug and play" capability when they are ready for global Internet connectivity. This important trend is seen, for example, in the metrics for IP registered addresses where we can note that less than one-third of the registered internets appear to have connectivity to the popular USA NSFNet backbone. Thus the announcement by Microsoft a few weeks ago at Interop that its enterprise networking strategy was based on TCP/IP, and the shipment the same week of the next generation Windows operating system NT with internet protocols "shipped in the box," will certainly have a dramatic effect as PCs all over the world reboot asking for their owners to feed them their Internet addresses. By analyzing the IP registration data base over the past six months, it has been possible for the first time to develop some comprehension where the near-term growth of IP internets is going. Clearly most of the growth - about 60 percent - is coming from commercial network registrations. In the USA, that figure is 70%. Other interesting trends include very fast growth rates of education internets, and significant increases in government networks. Overall, more than 2000 networks a month are now being registered by the RIPE European NIC, the new Asia-Pacific NIC, and the USA NIC. Internet Connectivity. One of the most meaningful internet metrics is that of connected networks. It's measure is imperfect because the best you can do is to look at the networks "seen" by the major backbones like the NSFNet or the CERNExchange. The whole conceptualization of the Internet - even in IP internet terms - was recast a few months ago by the discovery by a joint CERN- Merit team that the CERNExchange in the month of March saw 1 040 more networks than the NSFNet in the same period. As we have come to expect, the Internet continues to grow at a seemingly perennial rate of 10 percent per month - month after month, year after year. But, something new is occurring. What was largely a USA phenomenon has taken off worldwide. Not only is the current monthly growth rate of networks outside the USA 50 percent greater than in the USA, but also by June of 1994 at the present rates, most of the IP internets will exist outside the USA. Indeed, if the Bitnet and UUCP network infrastructures were considered in these figures, already the globalization crossover point has already occurred. Internet Hosts. In the final analysis, it's Internet hosts rather than networks that provide real information and people connectivity. Mark Lottor's famous quarterly "walks around the Internet" provide the benchmarks and trends for host connectivity. Like other metrics, it's still incomplete because it cannot reach hosts on non-IP internet infrastructures or hidden behind gateways. But it does provide a fairly good approximation of host connectivity. Perhaps the most interesting characteristic of this trend is its consistency over the past 12 years - as the number have just kept increasing from a handful to nearly 2 million machines. Whether you look long-term or short-term or in any region or country - the growth continues. Over the past 12-18 months, most notable increases from a globalization perspective have been developing countries in general, and Central & Eastern European countries in particular. Internet traffic. The last major metric category is internet traffic. As expected, with connected networks and hosts increasing at ten percent per month, the traffic is increasing globally as well. Major backbones like the NSFNet are currently transiting nearly 8 Terabytes per month and regularly jumping up another Terabyte. The commercial and European regional backbones seem to be humming along - just under a Terabyte per month. I think the global traffic remains a fascinating metric. Nearly every connected country engages in somewhere between one and one hundred Gigabytes of traffic per month to the NSFNet backbone. For nearly every country, the traffic growth has ranged between 800 and 2000 percent per year. Traffic through the CERN Exchange backbone has similarly risen dramatically. These statistics alone are a testimony to the significance of internetworking infrastructure throughout the world. Global communities of people and institutions are sharing ideas and information, they are collaborating on important professional, academic, social, and commercial developments - with a speed, ease, and scale that did not exist even two years ago. I think one of the best recent measures of this critical infrastructure realization is what is unfolding inside one of the old behemoths of the computer industry - IBM. Two years ago it began a crash program to provide Internet connectivity to every professional employee. It has been providing accounts to nearly 1000 employees a week, and in a recent survey, most professionals believed internet connectivity was important to their career. Internet Services. There is an unfortunate misunderstanding among the general public that internetworking is used just for pen pal kinds of activity or as a high-speed substitute for telex. Less known is that the Internet currently supports more than 1000 different kinds of computer services, although only about 400 are defined, registered services, and only a few dozen are in really widespread use. The rest are largely experimental or specialize gateways between different kinds of computer networks. Most of the traffic - around 50 percent - is file transfers. Internetworking allows extremely fast and easy movement of files, including "binary" files that constitute the software or special files used by computer programs. Not surprisingly, some of the Internet sites with the heaviest global traffic are those used for distributing new versions of software. The full scope of all the existing and emerging services today is worth listing: Basic Services - that include file transfers, e- mail and remote login to a distant computer. Mailing lists and bulletin board services that allow personal feeds of information. Interactive information delivery services that allow personal browsing. Services like Gopher, World Wide Web, and WAIS have been growing at annual rates exceeding one thousand percent Directory services that allow finding people Indexing services that allow finding information Active agents that allow automated gathering of information of which Knowbots are the most celebrated example. Network management to monitor and control networks and devices. Commercial electronic data that allow the exchange of commercial business records. The new Internet Merchantilism Initiative and Enterprise Integration network promise the rapid introduction of widespread interchange of commercial business data and forms. Applied encryption technology that allows property and monetary transactions and privacy enhanced mail. Multimedia capabilities that allow broadcasting and even entertainment Mobile capabilities that allow transparent connectivity on the move. The appearance of Personal Digital Assistants over the past few months and truly lightweight notepad PCs promises to fuel widespread mobile access. ***** What's driving the internetwork phenomenon? One of the most oft-asked questions today is "what's driving this revolution?" The answers are simple. TCP/IP protocols & applications came bundled with every Unix workstation - now virtually all the operating for PCs include the protocols. Internetworking coincided with enterprise internetworking revolution. Standards and applications were developed by the best standards process in the business with highly desirable characteristics such as speed, user driven, innovative, proven standards, openness, international scope, and rapid dissemination of the specifications and even code. User costs are very low due to "Sender keep all" accounting, inherently low-cost technology being employed, and institutional cost sharing. Massive global connectivity & professional use. Ability to directly reach any computer and process. Very high performance. ***** Who uses the internet? The second most asked question seems to be "who uses the Internet?" In short, nearly everyone: Institutions of all kinds - commercial, academic, and government - to allow their staff to collaborate with peers, to rapidly coordinate complex, dispersed worldwide activities; to gather and share information; by interconnecting their enterprise networks via Internet backbone providers Professional communities of all kinds - especially research and development organizations Business enterprises which specialize in providing or collecting information General public via local access providers and gateways to commercial public e-mail carriers and other kinds of networks ***** Why "critical information infrastructure"? One of the more interesting remarks made at the 1991 INET conference were those of the former deputy head of the USA White House Office of Science and Technology Policy, Gene Wong. He surprised people by stating that today, the USA was a developing country, for its institutions and people were just beginning the process of internetworking discovery and development - together with much of the rest of the world. Although there has indeed been considerable focus under the new USA administration on internetworking as critical national information infrastructure over the past few months; perhaps even more significant were the findings and initiatives announced by the World Bank at its first internetworking seminar three weeks ago. Drawing on the World Bank findings - as well as the experiences of the past few years, it's possible to list specific factors that indeed make internetworking critical not only to countries, but to institutions and professional and social activities as well. Scalability and robustness. One of the more remarkable characteristics of internetworking is the ability to use so many different technologies and platforms at all different levels - transport, network, and application. You can use whatever exists or is economical. In some places that means low-end DOS or Unix boxes using dialup lines for Fidonets or UUCP networks. It means Bitnets were there's an existing infrastructure of IBM machines. Or it scales to the high end with FDDI LANS, high performance routers and DS3 or ATM pipes. One of the world's great unsung testimonies to the inventive genius of people yearning to internetwork is what our friends and colleagues in Central and Eastern Europe and Russia have done to done to assemble vast functional internetworks that squeeze every bit from the available infrastructure. Unfortunately, internetworking's robustness in the adaptive routing domain has even caused futile export control laws to be applied which are hopefully now becoming a thing of the past. Leapfrogging. Internetworking provides a way to work with and work around existing infrastructure. There is nothing that can intrinsically retard the introduction of internetworking technologies and applications - - except for unenlightened authorities that impose artificial high cost cost, standards, or export barriers. Other Infrastructure Support. Internet- working is the means for countries to much better support the growth and development of nearly all other basic national infrastructures - a phenomenon noted in recent World Bank studies. Access to Resources. Internetworking provides instantaneous access to resources - which include not only an estimated several Terabytes of information, but also an enormous numbers of specialists in different professional fields. For example, a few months ago, a plea by the World Health Organization on Internet global medical discussion groups for someone familiar with the unusual symptoms of a little girl produced a tropical disease specialist who produced a diagnosis and recommended medication that possibly saved the child's life. Critical Skills. Internetworking - in a kind of self-propagating manner - allows large numbers of computer, networking, and telecommunications students and business personnel to acquire the skills and familiarity with the technology to allow its introduction and proliferation on a wide scale. It also allows specialized professionals and researchers in all fields to "stay home" because they can practice their profession and interact with colleagues worldwide on the network. This factor was, for example, one of the prime motivations behind the Australian AARNET initiative to connect every corner of that vast continent. Business Orientation. One of the more significant factors mentioned by the World Bank was the importance of having the necessary infrastructure and orientation to support a market-oriented economy. Internetworking accomplishes this rather admirably, and it's no secret that most of the world's banks and financial institutions not only have their own internal internetworks, but are increasingly using the world's Internet mesh to do business as well. Funding Sources. The increasingly recognized importance of internetworking has led a wide variety of agencies and foundations and even private companies around the world to fund many diverse internetworking initiatives and infrastructures - domestically and internationally. France, Italy, the CEC, and the USA have, for example, been aggressive in pursuing major international initiatives. The donors for this conference - - of course - are particularly notable examples of this trend! Motivated Community. And lastly, the internetworking community must be one of the most enthusiastic and devoted in the world. So many of you here are representative of this amazing global phenomenon that has resulted in so many people spending so many hours and so much creative energy in scaling the availability and use of these technologies. Why? I suppose that it's because internetworking is dramatically changing the world in so many ways, because it is both business and fun at the same time, and because it is possible to see rather quickly the tangible benefits of one's own individual endeavours. ***** Where are we heading - An Internetwork Renaissance Based on every indicator, continued exponential movement can be expected along what I call the three major axes of internetworking infrastructure: performance, ubiquity, and application. On the ubiquity front, it is certainly fun to contemplate John Quarterman's famous extrapolations of Internet users and the human population and fantasize about their meeting in the year 2001 in a world where everyone and every human activity are networked. That's not likely to happen, however, even if internetworking may become very extensive. Access to all forms of electronic communication is pretty much a function of GNP and relative need - and mere survival is still a problem for much of the world. On the other hand, as the World Bank noted, there are significant societal support systems that are dramatically enhanced by internet connectivity, and those are beginning to find their way into many remote corners of the world. With the introduction of simple high-speed dialup nationwide and even worldwide "free- phone" internet services (which I'm actually beta-testing), will significantly scale ubiquitous access. Other indicators are the continuing massive institutional connectivity of businesses, government agencies, and schools; the appearance of the "MacDonalds" style local Internet kiosks; the converters that turn Cable TV systems into Ethernets; and the satellite systems that will allow so many 3rd and 4th world countries to continue to connect. On the performance front, already there are experimental implementations of IP running at near Gigabit per second rates, as well as over ATM. Increasing liberalization and competition worldwide of telecoms should begin forcing the availability of raw leased circuit bandwidth from stratospheric levels down toward cost. And lastly but most significantly on the application front, the availability of internetworking tools on mass market PC platforms combined with the massive professional and institutional connectivity should produce the equivalent of an Internet- working Renaissance the likes of which the world has never witnessed! *****