If you were going to design the next version of the Internet, how would you go about it?
This is the question that a combination of government, academic, and industry groups are trying to answer by crafting several new, high-speed networks, such as Internet2 and Next Generation Internet. In its first generation, the Internet was tightly controlled by the Department of Defense, until the National Science Foundation came along and extended it to university supercomputer centers and eventually the academic community at large.
But with this newest iteration, diverse interests ranging from the National Science Foundation to Microsoft to IBM to the University of Hawaii to Al Gore are involved. Perhaps now the question ought to be: Who is this next Internet for, and who should foot the bill?
In the short term, at least, the answer appears to once again fall on the federal government and public and private universities.
"The federal government brings some essential seed money where commercial interests - necessarily driven by shareholder responsibilities - don't yet see a proven concept but do see too much of a risk to make large investments alone," said Heather Boyles, chief of staff for the Internet2 project.
Thus far, the Clinton administration has budgeted US$100 million annually for five years to the White House's NGI initiative. The National Science Foundation, too, is contributing grant money to high-speed Internet projects, like the university-led Internet2. Meanwhile, the industry that has capitalized on the current Internet - the Ciscos, Microsofts, and Oracles of the world - have made comparatively minor cash investments in such projects though they do help implement the technology.
Given the breakneck pace at which the computing industry has managed to capitalize on the Net, it is perhaps surprising that so many feel that universities are still a key part of the process.
"If we didn't have a public/private partnership, you would see slower, more incremental changes, and it would take longer to implement next-generation capabilities," said George Strawn, NSF's division director for networking and communications research and infrastructure. "Industry has its hands full catching up with demand and providing services, which is a wholly occupying activity."
William Graves, an Internet2 steering committee member and University of North Carolina provost, said the evolution of the Internet2 project - as well as past and subsequent Internets - occurs in a four-stage cycle: research, incubation, productization, and commoditization. Some researchers argue that the government's investment in new Internet projects is tiny compared to the technological and economic impact those investments have.
"The reason the US is so far ahead [in Internet technology] is because of this interaction between the federal government, universities and industry," said Larry Smarr, director of the National Center for Supercomputing Applications at the University of Illinois. "The federal government is spending a few tens of millions of dollars which causes a collective movement of billions of dollars from the public and private sector every year. It's rare that a government action can cause such a social phase transition in universities," he explained
In addition to having the need for more advanced research and teaching applications, universities have a ready supply of guinea pigs - otherwise known as students - that happily develop and test new technologies. However, the issue of students developing new, marketable Internet technologies has prompted some universities to create new intellectual property rights policies to protect their investments, which could create more and more conflicts of interest in the future.
When Netscape's Marc Andreessen left the University of Illinois with the code for the Mosaic browser, university lawyers and the National Center for Supercomputing Applications - where the browser project was developed and incubated - threatened a lawsuit over the Mosaic name and Andreessen's right to commercialize a university-led project. But for beta-testing new technologies, it's hard to find a more interested and captive audience.
The Internet2's transition from testbed to industry to society at large will take about four or five years, said Graves and other project leaders. In the meantime, university researchers will be testing heavy-duty applications that take advantage of new technologies like "Quality-of-Service" guarantees, which will allow applications to request a specific amount of bandwidth or priority for real-time interactions.
The NSF's Strawn said new Internet projects will be funded with government money as long as there is a perceived need, and though the boundaries between projects are fuzzy, they have mutually supportive and compatible goals.
"As part of our research," said John Mansfield, an associate research scientist at the electron microbeam analysis lab at the University of Michigan, "we'll put up with the glitches that the industry people won't, and what the NSF is paying us to do is work out the bugs and research it."
