|
[Date Prev]
[Date Next]
[Thread Prev]
[Thread Next] Internet Railroad
Tom - A short technical note on the Internet Railroad. We're really encouraged by the Administration's support and the offer to combine the Internet 1996 World Exposition with the application areas decided at the G-7 summit. Combining the two sets of requirements is very doable, but will have to be done carefully so the technical objectives are met. I'm hoping this note will give a bit more detail about how the Railroad works and how we can combine this effort with the G-7 application areas. The Internet Railroad carries "backend" traffic such as managed video streams, FTP mirrors, Web mirrors, or other "bulk" tranfers of data. The railroad does not carry general-purpose transit traffic. This is only for public and group categories of data that are part of the Internet 1996 World Exposition. (Some of our participating institutions include the Kennedy Center, the National Press Club, and the Tokyo Aquarium. Full details on the fair are at http://town.hall.org/fair.) The Internet Railroad we envision has a 45 million bit per second (or faster) core connecting up to seven of the key interconnect points on the Internet. At each "stop" there will be a cluster of our equipment. This includes a large server (Quantum is providing 1 terrabyte of data plus we have significant donations in place for other components). There will also be a router, that is used to connect our railroad to the general purpose Internet infrastructure. The interconnect to the Internet is a crucial point. One of the purposes of the fair is to stimulate new home markets for the Global Information Infrastructure. By placing the stops of our railroad at the major interconnect points, it means that commercial providers can distribute (e.g., sell) the data coming off the railroad to their consumer bases. One of the major interconnect points in the U.S. is in Washington, D.C., known as MAE-EAST (Metropolitan Area Exchange, Eastern U.S.). This 100 mbps fiber network is used to interconnect many of the world's largest networks include Alternet, PSI, MCI, Sprint and the NSF's International Connection Manager (ICM) program. In Tokyo, there is a similar interconnect point where the Japanese commercial and research networks all exchange traffic. On the West coast of the U.S., the NASA Ames Research Center provides the MAE-West interconnect point. Finally, there are similar locations in Amsterdam and London. We have agreements in place with all the organizations to allow the World Exposition facilities to be located directly on the interconnect points. It is crucial that the Internet Railroad connect these main points. In addition to the core 45 mbps of the railroad, we are trying to establish lower-speed links to regional centers. For example, a T1 from Tokyo to Bangkok is being installed to extend the railroad to the Thailand Internet Exchange. From Amsterdam, we will be using NATO-provided satellite transmission to reach Central Asian republics such as Armenia or Georgia. There are two crucial points for us. First, the location of any transoceanic links is crucial. Connecting the key exchange points is what will make this a real world's fair for the information age. In particular, we'd really like to connect NASA Ames Research Center to Tokyo and to connect Washington, D.C. to London and Amsterdam (and Stockholm if possible). Second, it is very important that we look at the type of bandwidth carefully. We'll have a very robust network infrastructure with considerable equipment and a very large Network Operating Center. We'd love to support the applications on our infrastructure and believe we can demonstrate a very credible case for being able to do this well: our engineering team includes leading research groups from all over the world and their is a very strong committment to providing a solid operating base for fair activities as well as the application demonstrations. Another possibility is that the bandwith be multiplexed into several smaller links, allowing the leading-edge applications to experiment with technologies such as ATM or direct Sonet interfaces without an undue impact on fair operations. I think its important that we look at a variety of goals for 1996. Some of the goals are to demonstrate interoperability and advance the state of the art. At the same time, we're also trying to spur new markets and provide a very visible demonstration of the power of information technology (a classic function of a world's fair). Your assistance in making the World Exposition a part of the 1996 GII activities has been very helpful. I'm looking forward to getting some resource committments in place and figuring out how to use all those bits! Carl Malamud Internet Multicasting Service
|