This story appeared on Network World at


Internet architecture: How distant are the elephants?


'Net Insider 


By Scott Bradner, Network World, 08/21/06


In 1992 Dave Clark, an MIT researcher and the original Internet architect, exhorted the IETF to deal with "distant elephants," such as security and addressing, and their impact on the Internet. He reprised the same talk more than 10 years later during the IETF's 20th anniversary meeting in Dallas. Now, Clark and others are using our tax dollars to explore seriously visions of how a new Internet should work.


Clark's original talk (see page 539) has aged well. About the only elephant he discussed that turned out to be vaporware is ATM - people were more willing to connect to an insecure Internet than Clark thought they might be. But overall, the talk is as relevant today as it was in 1992, and Clark is singing much the same tune.


Clark's current work is explored in "The Internet Is Broken," a special report in MIT's Technology Review. Don't be put off by the magazine's overhyped teaser for the report: "The Net's basic flaws cost firms billions, impede innovation, and threaten national security. It's time for a clean-slate [approach]." The report itself is better than the teaser.


In the report - and live - Clark is pessimistic about the Internet of today. He worries "we might just be at the point where the utility of the Internet stalls - and perhaps turns downward." As he was in 1992 when he said, "lack of security means the end of life as we know it," Clark is focused on the lack of security functions within the Internet. This lack means that security is the responsibility of the user, someone who is unlikely to be a security expert. Security is not the only issue but is painted as the key one.


Clark's four goals for a new Internet architecture are these: There should be a basic security architecture that includes authentication of Internet users; ISPs should be enabled to offer advanced services without compromising their businesses; devices of all sizes should be able to connect to the Internet; network management should be easier and more resilient.


The U.S. National Science Foundation is putting some of our tax money into researching a new Internet architecture. It has created the Global Environment for Networking Innovations (GENI) Initiative to "explore new network capabilities that will advance science and stimulate innovation and economic growth." In the past, NSF-sponsored research was key to the development of today's Internet, and I hope that the GENI Initiative's research will enable tomorrow's better, safer, more economically viable and more manageable Internet.


The GENI Initiative is not the only game when it comes to designing an Internet for tomorrow. For the past few years, the International Telecommunication Union has been working on the Next Generation Network (NGN) Global Standards Initiative.


The ITU's goals for its network sound quite a bit like Clark's, but the players are quite different. GENI works with network researchers, and the NGN is mostly a telephone-industry effort.


Some players envision the NSF's GENI and the ITU's NGN as new infrastructures running in parallel to today's Internet. This reminds me of the dream fans of ATM once had. My midyear prediction is that the Internet you will be using 10 years from now will have technologies that came from both of these initiatives but the Internet of tomorrow will be closer to today's Internet than either of the initiatives expect.


Disclaimer: There are a lot of educational infrastructures that run in parallel to Harvard's - and it's up to you to judge their relative merits - but the above worry about parallelism is my own.




All contents copyright 1995-2006 Network World, Inc.