Frustrated America Online subscribers who couldn't dial on to that network last week aren't the only people to find themselves the victims of online overcrowding. Among the most impatient users of computer networks are the scientists and academics for whom the Internet was originally created. It has become, they say, too crowded, too slow and too primitive. And they want a brand new Internet all their own.
A formal proposal, informally dubbed "Internet II," was launched last October, backed by 100 universities and research organizations who pledged $500,000 a year to help build a new network over the next three to five years in conjunction with commercial companies and government agencies. MCI, using existing fiber-optic lines, is involved. Cisco Systems and IBM have pledged their support. And this spring the White House will ask Congress for $100 million annually to help finance the project as part of a government proposal called the Next-Generation Internet Initiative.
Salon spoke recently with Tom Kalil, the Clinton administration's technology policy chief, about Internet II and the government's role in it.
What are the key goals of the Next Generation Internet Initiative?
From the administration's point of view, there are three. The first has to do with bandwidth: How do we get 100 universities connected to the Internet at 100 times the current speed, and a small number of other sites connected at 1,000 times the current speed? The second goal is investing in networking technology like resource reservation, multi-casting and scaling so that the Internet will not only be 100 times faster but have 100 times more devices attached to the network. It is not something an ordinary end-user would interact with, but it is absolutely essential to the operation of the network.
The third goal is to demonstrate a series of applications called distributive computing and to set up "collaboratories," which will allow researchers around the country to interact as if they were in the same building.
The idea isn't just to make AOL's life -- and that of the average Internet user -- easier?
No, we're not interested in funding commodity research. With Internet II, we're talking about an architecture that would allow universities and other agencies to separate advanced research traffic that really needs high bandwidth from the commodity traffic that they can get from an Internet service provider. This is not about going back. This is about allowing some subset of the research community to look at the future.
What is wrong with the Internet we have now?
Quality of service. Right now the Internet only offers one quality of service, called "best efforts." If I send you an e-mail it gets there on a first-in, first-out basis. Well, that's fine for e-mail because I don't really care whether you get it in 300 milliseconds or a second. But if I'm trying to have a real-time, 30 frame-per-second video conference, then the delay and jitter that you get over today's Internet makes for an unsatisfactory experience.
We also think there are a bunch of research issues involved in scaling the Net. We not only want to attach computers to the Internet but also thermostats, microwave ovens or cell phones, for example. So you could send your thermostat an e-mail message saying, "I'm coming home in 20 minutes and I'd like it to be warm by the time that I get there."
The government, mostly in the form of the Defense Department, was central to the first Internet. But with the private sector now so heavily involved, does the government still need to have a role?
One of the reasons the government is working on Internet II is because the research community and government agencies have some very demanding applications. For example, next year a NASA project called Mission Planet Earth will launch satellites that will generate trillions of bytes of information. You could not FTP that over today's Internet. So if NASA wants the research community to be able to interact with it in real time, it needs superior services. The same with the Department of Energy if it wants to access the Lawrence Berkeley Laboratory Advanced Light Source from their desktops.
You're asking for $100 million, which would begin to kick in next October. But given the budget-cutting sensibilities and the fractious political climate in Congress, how realistic is that figure?
We'll have to make a case for it. Looking at it strictly from an economic point of view, it makes a great deal of sense. The Wall Street Journal recently published an analysis of how much wealth was being created by the Internet, and the figure they came up with was $250 billion. The money would allow U.S. universities and companies to develop new research prototypes of new products and services. It will give U.S. firms a big leg up in international markets in the future.
$100 million does not seem like a lot of money for such an ambitious project.
It's not just federal government money. Universities will each be putting up a half million of their own money each year. There will be matching funds from private industry. So there could be hundreds of millions of dollars per year for this project. This project creates an excuse for people who want to do this type of research to run to the university president and say, "Look, this is a national Internet initiative announced by the president, and our university has to be involved. I mean, compare the costs of this to a new football stadium!" So with the Next Generation Initiative, we're running up the flag and people are starting to salute.
How do you see the relationship between the government and private universities as the initiative develops?
Our strategy is aimed at giving the money to the universities and allowing them to purchase commercial services as they become available. This is different than the way we did things in the past, when the emphasis was on subsidizing a network. Now, what we're trying to do is increase the purchasing power of the universities so that they can go out and buy really high-end services as they become commercially available.
The National Science Foundation is the lead government agency in this project.
Yes. The NSF already has a program up and running as part of the initiative called High Performance Connections, a very high-speed backbone network service connected to five supercomputer centers across the U.S. NSF grants will allow a greater number of universities to connect to this backbone.
Who determines how that money is spent?
The way it goes with the High Performance Connections program is that universities tell us what sort of applications they want to do. They don't say: "Give me some money for an OC3 line." They say: "These are the applications that we want to do and this is why we need them."
But overall, given the limited resources -- at least, initially -- who decides which applications get the most money or which of them will get money first?
On the basis of a competitive, merit-based review.
And who chooses the winner?
Whoever issues the request for proposals. If it's NSF, then it will be NSF. If it's DARPA (Defense Advanced Research Projects Agency) then it will be DARPA. Other agencies involved in this will be more mission-oriented agencies like the Energy Department and NASA. But how that $100 million will get divided up among the agencies hasn't been decided yet.
Who said the era of big government was over?
I think that sometimes there's an inadequate appreciation of the role that government research has historically played. Everyone says, "The Internet is a paragon of private sector enterprise." And that's exactly right. But what they forget is the government started supporting this back in 1969. And it wasn't until there was a very active academic network that the commercial sector said, "This market is now big enough for us to actually get in."
Bullets or ballots?
"We need to develop normal ties so that we won't have to wage war anymore."
-- Aslan Maskhadov, the guerrilla commander considered the front-runner in Monday's presidential elections in Chechnya. (From: "Chechens cast ballots in national elections," reported Monday by Cable News Network.)