📄 rfc1336.txt
字号:
Nevertheless, with growth and success (plus subtle policy shifts in Washington), the prevailing mode may be shifting towards competition, both commercial and academic. To develop protocols in a commercially competitive world, you need elaborate committee structures and rules. The action then shifts to the large companies, away from small companies and universities. In an academically competitive world, you don't develop any (useful) protocols; you get 6 different protocols for the same objective, each with its research paper (which is the "real" output). This results in efficient production of research papers, but it may not result in the kind of intellectual consensus necessary to create good and useful communication protocols. Being a member of the IAB is sometimes very frustrating. For some years now we have been painfully aware of the scalingMalkin [Page 5] RFC 1336 Who's Who May 1992 problems of the Internet, and since 1982 have lived through a series of mini-disasters as various limits have been exceeded. We have been saying that "getting big" is probably a more urgent (and perhaps more difficult) research problem than "getting fast", but it seems difficult to persuade people of the importance of launching the kind of research program we think is necessary to learn how to deal with Internet growth. It is very hard to figure out when the exponential growth is likely to stop, or when, if ever, the fundamental architectural model of the Internet will be so out of kilter with reality that it will cease be useful. Ask me again in ten years. 4.3 Hans-Werner Braun, IAB Member Hans-Werner Braun joined the San Diego Supercomputer Center as a Principal Scientist in January 1991. In his initial major responsibility as Co-Principal Investigator of, and Executive Committee member on the CASA gigabit network research project he is working on networking efforts beyond the problems of todays computer networking infrastructure. Between April 1983 and January 1991 he worked at the University of Michigan and focused on operational infrastructure for the Merit Computer Network and the University of Michigan's Information Technology Division. Starting out with the networking infrastructure within the State of Michigan he started to investigate into TCP/IP protocols and became very involved in the early stages of the NSFNET networking efforts. He was Principal Investigator on the NSFNET backbone project since the NSFNET award went to Merit in November 1987 and managed Merit's Internet Engineering group. Between April 1978 and April 1983 Hans- Werner Braun worked at the Regional Computing Center of the University of Cologne in West Germany on network engineering responsibilities for the regional and local network. In March 1978 Hans-Werner Braun graduated in West Germany and holds a Diploma in Engineering with a major in Information Processing. He is a member of the Association of Computing Machinery (ACM) and its Special Interest Group on Communications, the Institute of Electrical and Electronical Engineers (IEEE) as well as the IEEE Computer Society and the IEEE Communications Society and the American Association for the Advancement of Science. He was a member of the National Science Foundation's Network Program Advisory Group (NPAG) and in particular its Technical Committee (NPAG-TC) betweenMalkin [Page 6] RFC 1336 Who's Who May 1992 November 1986 and late 1987, at which time the NPAG got resolved. He also chaired the Technical Committee of the National Science Foundation's Network Program Advisory Group (NPAG-TC) starting in February 1987. Prior to the organizational change of the JvNCnet he participated in the JvNCnet Network Technical Advisory Committee (NTAC) of the John von Neumann National Supercomputer Center. While working as Principal Investigator on the NSFNET project at Merit, he chaired the NSFNET Network Technical Committee, created to aid Merit with the NSFNET project. Hans-Werner Braun is a member of the Engineering Planning Group of the Federal Networking Council (FEPG) since its beginnings in early 1989, a member of the Internet Activities Board (IAB), the Internet Engineering Task Force. He had participated in an earlier, informal, version of the Internet Engineering Steering Group and the then existing Internet Architecture Task Force. While at Merit, Hans-Werner Braun was also Principal Investigator on NSF projects for the "Implementation and Management of Improved Connectivity Between NSFNET and CA*net" and for "Coordinating Routing for the NSFNET," the latter at the time of the old 56kbps NSFNET backbone network that he was quite intimately involved with. ------------ The growth of the Internet can be measured in many ways and I can only try to find some examples. o Network number counts There were days where being "connected to net 10" was the Greatest Thing Ever. A time where the Internet just consisted of a few networks centered around the ARPAnet and where growing above 100 network numbers seemed excessive. Todays number of networks in the global infrastructure exceeds 2000 connected networks, and many more if isolated network islands get included. o Traffic growth The Internet has undergone a dramatic increase in traffic over the last few years. The NSFNET backbone can be used as an example here, where in August 1988 about 194 million packets got injected into the network, which had increased to about 396 million packets per month by the end of the year, to reach about 4.8 billion packets in December 1990. January 1991 yielded close to 5.9 billion packets as sent into the NSFNET backbone.Malkin [Page 7] RFC 1336 Who's Who May 1992 o Internet Engineering Task Force participation The early IETF, after it spun off the old GADS, included about 20 or so people. I remember a meeting a few people had with Mike Corrigan several years ago. Mike then chaired the IETF before Phill Gross became chair and the discussion was had about permitting the "NSFNET crowd" to join the IETF. Mike finally agreed and the IETF started to explode in size, now including many working groups and several hundred members, including vendors and phone companies. o International infrastructure At some point of time the Internet was centric around the US with very little international connectivity. The international connectivity was for network research purposes, just like the US domestic component at that point of time. Today's Internet stretches to so many countries that it can be considered close to global in scope, in particular as more and more international connections to, as well as Internet infrastructure within, other countries are happening. o References in trade journals Many trade journals just a year or two ago had close to no mention of the Internet. Today references to the Internet appear in many journals and press releases from a variety of places. o Articles in professional papers Publications like ACM SIGCOMM show increased interest for Internet related professional papers, compared to a few years ago. Also the publication rate of the Request For Comments (RFC) series is quite impressive. o Congressional and Senatorial visibility A few years ago the Internet was "just a research project." Today's dramatically increased visibility in result of the Internet success allows Congress as well as Senators to play lead roles in pushing the National Research and Education Network (NREN) agenda forward, which is also fostered by the executive branch. In the context of the US federal government the real credit should go to DARPA, though, for starting to prototype advanced networking, leading to the Internet about twenty years ago and over time opening it up more and more to the science and research community until more operationalMalkin [Page 8] RFC 1336 Who's Who May 1992 efforts were able to move the network to a real infrastructure in support of science, research and education at large. This really allowed NSF to make NSFNET happen. o Funding The Internet funding initially consisted of DARPA efforts. Agencies like NSF, NASA, DOE and others started to make major contributions later. Industrial participation helped moving the network forward as well. Very major investments have been made by campuses and research institutions to create local infrastructure. Operational infrastructure comes at a high cost, especially if ubiquity, robustness and high performance are required. o Research and continued development The Internet has matured from a network research oriented environment to an operational infrastructure supporting research, science and education at large. However, even though for many people the Internet is an environment supporting their day-to-day work, the Internet at its current level of technology is supported by a culture of people that cooperates in a largely non-competitive environment. Many times already the size of the routing tables or the amount of traffic or the insufficiency of routing exchange protocols, just to name examples, have broken connectivity with many people being interrupted in their day-to-day work. Global Internet management and problem resolution further hamper fast recovery from certain incidents. It is unproven that the current technology will survive in a competitive but unregulated environment, with uncoordinated routing policies and global network management being just two of the major issues here. Furthermore, while frequently comments are being made where the publicly available monthly increases in traffic figures would not justify moving to T3 or even gigabit per second networks, it should be pointed out that monthly figures are very macroscopic views. Much of the Internet traffic is very bursty and we have frequently seen an onslaught of traffic towards backbone nodes if one looks at it over fairly short intervals of time. For example, for specific applications that, perhaps in real-time, require an occasional exchange of massive amounts of data. It is important that we are prepared for more widespread use of such applications, once people are able to use things more sophisticated than Telnet, FTP and SMTP. I am not sure whether the amount of research and development efforts on the Internet has increased over time, less even kept pace withMalkin [Page 9] RFC 1336 Who's Who May 1992 the general Internet growth (by whatever definition). I do not believe that the Internet is a finished product at this point of time and there is a lot of room for further evolution. 4.4 Ross Callon Ross Callon is a member of the Distributed Systems Architecture staff at Digital Equipment Corporation in Littleton Massachusetts. He is working on issues related to OSI -- TCP/IP interoperation and introduction of OSI in the Internet. He is the author of the Integrated IS-IS protocol (RFC 1195). He has also worked on scaling of routing and addressing to very large Internets, and is co-author of the⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -