Posted at Friday, May 18, 2007  EDT  

Untangling the World Wide Web

CHRISTOPHER DREHER,  From Saturday's Globe and Mail

For the past few decades, a huge network infrastructure has provided billions of people with access to information and technology that was inconceivable to earlier generations.

But if the cybergelicals of the 1990s were right about how the Internet would transform everyday life, they were less prophetic about what exactly those transformations would look like. As the number of users and applications has expanded, so have the frustrations and risks of plugging in.

This week, the United States banned soldiers from websites such as YouTube and MySpace – concerned that downloads and social networking could overload military systems and lead to security breaches. Some banks have reverted to snail mail to help customers steer clear of phishers trying to bilk them out of their money.

“Over all, the situation is not getting better, it's getting worse,” says David Clark, a senior researcher at the Massachusetts Institute of Technology and one of the early creators of the Internet.

He points out that thousands of researchers have been trying to remedy Internet security issues, among other flaws, for decades. “We're not going to get there by trying harder, but by trying different.”

Which is why a growing number of programmers and network specialists – increasingly skeptical about jury-rigging solutions for today's sprawling, technically discordant Internet – are asking a seemingly heretical question: Should we throw out the Web and start over?

Last month, a group of computer scientists at Stanford University in California formally launched a research program called the Clean Slate Design for the Internet Project. Their goal is to reimagine the Web's basic architecture.

“It's a fundamental change in thinking,” says Nick McKeown, an associate professor at Stanford who directs research at Clean Slate. “Instead of trying to fix problems for today, we're trying to figure out what the Internet should look like in 15 years.”

THE SERVER IS BUSY

The progenitor of the Internet was first launched in the 1960s by the U.S. Department of Defence. Concerned about communications in the event of a crippling disaster, such as a nuclear attack, they funded experiments in the military applications of network technology.

The National Science Foundation also saw the potential of networking. By the 1980s, their researchers rapidly expanded the technology in the hopes of connecting academics and scientists across the country – eventually developing a fluid, egalitarian system that would form the backbone of the Web as we know it.

As now-popular mythology has it, the democratic – almost anti-authoritarian – nature of burgeoning Internet culture also helped it make the leap from scholars trading thesis drafts to mainstream use. And ultimately deepened the potential of the network.

But ironically the Internet's openness is also its albatross. Because of ad-hoc innovations, the Web has become a kind of unwieldy trailer park of technology – where security and even fundamental stability remain highly problematic.

For example, early adopters and researchers weren't worried about an interminable stream of e-mail from strangers hawking Viagra or spammers posing as eBay asking for personal financial data, so security has been developed in a patchwork manner.

The Internet was not designed for Second Life or “adult entertainment” videos either – high-volume, resource-consuming uses of the network. If just 1 per cent of the DVDs that NetFlicks sends to customers every day were downloaded, we would need a tenfold increase in the current core capacity of the Internet.

The Internet was also created for static, wired computer-to-computer communication, not for the burgeoning demand for mobile connectivity. While people now get e-mail and browse the Web on PDAs and cellphones, that online connection is being tunnelled through cellphone network providers, which offer a much smaller bandwidth capacity.

All of this leaves today's geeks with daunting problems and few easy fixes. Which might explain why most have stuck to the same basic assumptions about the Internet as early founders – simply propping up the old network as they go.

“In every other high-tech field, it's usually typical to see massive innovation,” Prof. McKeown says. “And although we've seen huge implementation of new applications, Internet technology is built on the same ideas it was built on 40 years ago.”

Still, after you get a flat tire, you can drive on the rim for only so long before you have to pull over, or risk worse damage. This is why Guru Parulkar has long advocated the “clean slate” solution. Currently the director of programming for the National Science Foundation in Arlington, Va., he believes that the Internet has become dangerously “ossified.”

In August, he will take the helm of Stanford's Clean Slate Design. The ideas researchers come up with there will then be tested on a beta network of sorts he spearheaded at the NSF called the Global Environment for Network Innovations. It will cost $300-million to $400-million (U.S.).

STARTING FROM SCRATCH

Although the work at Clean Slate involves highly technical considerations – such as a redesign of the wireless spectrum allocation to better use limited network capacity – its success could greatly affect our daily lives.

Better wireless spectrum allocation, for instance, would finally mean faster and more foolproof data communication between handheld wireless devices such as phones and PDAs. It would also fulfill at last the promises of devices that combine the capacities of a television, a DVD player and a home computer.

Likewise, improving network security would mean that instead of spending billions of dollars preventing spam, virus attacks, malicious hacking and other dangers, businesses could expand on some of the life-altering real-time uses imagined by pioneers of the Internet.

Remote surgery, for instance, has been performed on a very limited basis since its first success in Canada in 2001. But it can take place only over dedicated fibre-optic cables because the Internet networks used by the general public have too many unforeseen variables, including security concerns and possible blips in connectivity.

These issues also prevent a range of other industries and many critical infrastructures – such as water and electric plants or airports and highways – from fully using the Internet. “If air-traffic control were run over the public Internet,” Prof. McKeown says of the current system, “then I wouldn't fly.”

A “clean slate” Internet might also open up what Mr. Parulkar calls “the interaction between the physical and virtual worlds.” For example, PDAs could communicate with everyday objects embedded with chips or sensors that might make window shopping literal. When you drive by a store, your PDA could display sale prices on that leather jacket you've been eyeing.

Of course, Mr. Parulkar says we would need a network 100,000 times bigger than what we have now to handle this kind of traffic.

Yet despite the seemingly impenetrable technological aspects of overhauling the Internet, many of the most difficult issues have nothing to do with circuits or wires or bandwidth. “For us, the easiest questions to answer are the purely technical ones,” Prof. McKeown insists.

The real hitch? Ask telecommunications companies such as Bell and AT&T, which became Internet providers in the mid-1990s in the hopes of making huge fortunes. “One of the dirty little secrets of the network is that the network infrastructure is not economically sustainable or profitable,” Prof. McKeown says.

In fact, he wonders if the only economically sustainable model for the Internet may be a nationally funded or regulated infrastructure – or some sort of government monopoly. (Though he adds that, “in the current economic and political climate” of the U.S., proposing this idea “is nearly suicide.”)

Another thorny issue facing advocates of a “clean slate” approach to the Internet is how to balance privacy and security concerns. Making the network less open to spam and viruses, for example, also means curtailing the freedom and anonymity of the Internet.

This is why Stanford's program will be bringing together a range of experts from different fields – programmers and economists, to start – to examine the roles that policy, funding and government should play in a new conception of the Internet.

Meanwhile, the ultimate impact of projects such as Clean Slate remain unclear. Advocates don't believe it will replace the current Web entirely. Perhaps a new network will simply create a parallel system that can run in tandem with what we have now.

“No one could have predicted that the Web would come along,” Prof. McKeown says. “And the same type of unforeseeable thing could happen. Those of us who have been on the Internet for a long time believe that we know the right things, but we could be wrong.”

New York writer Christopher Dreher is a frequent contributor to The Globe and Mail's Focus section.