Warning Signs on the Net

Within the past two weeks, several military experts — including no less an authority than Gen. John P. Abizaid, the senior U.S. commander in the Middle East — have described the U.S. military’s core challenge in Iraq as essentially one of information. There are probably 5,000 or fewer armed guerrillas who oppose some 155,000 coalition troops, Gen. Abizaid said. The problem is getting information about whom and where those guerrillas are.

This situation stands in interesting contrast to the initial spring invasion. Then, America’s victory came in part because the problem of gathering, transmitting, and analyzing information about the combatants was solved by IT hardware and software, as InfoWorld pointed out in its June 2 story “From the battlefield to the enterprise.” Today the problem is more human and political than technological. And no amount of technology can — by itself — assure victory.

The situation serves as a reminder of the limits of IT, as does this week’s brilliant story, “The threat to universal Internet connectivity,” by Senior Contributing Editor David Margulius. From its earliest days as a network for researchers allied with the Department of Defense, the Internet was designed to have as few moving parts as possible. Architects such as Vint Cerf, who was among those Margulius interviewed, kept the core of the network very simple.

The aim was partly to make the network more durable — as a Cold War invention, it had to survive a nuclear attack. But the philosophy was also to encourage innovation at the Net’s periphery, where most of us intersect with it. Since then, the Net has grown beyond even its founders’ wildest dreams. And now a growing number of sober revisionists, including some of the stewards of the network, believe there must be more intelligence built into the core to cope with this congestion.

The issue came to a head a couple of months ago when VeriSign, which runs the DNS, altered that worldwide mechanism for looking up the numerical IP address of domain names such as infoworld.com. The change: When a person mistyped a name (e.g., “inofworld.com”), they were rerouted to a VeriSign search site rather than getting the usual error message.

Unfortunately, the change wreaked havoc with existing software —  spam filters, e-mail applications — that relay error messages. VeriSign withdrew the change under pressure a few weeks later, although CEO Stratton Sclavos[ maintains that some such alteration will need to be made to ensure capacity and security.

The real problem, of course, is that no one is in charge of the Internet. It’s managed by a loose coalition of industry, government, academic, and standards-setting groups, and there isn’t a strong central forum for resolving such disputes. And with international barriers increasingly being erected — China, for example, routinely blocks its citizens from accessing certain sites — there is a very real danger that the Net will become divided into a series of camps without true or reliable end-to-end connectivity.

Here again, as in Iraq, the problem is not technological but human and political — how to get these disparate groups to cooperate. While Margulius offers not pat solutions, his is the most comprehensive analysis of the dangers and opportunities I’ve seen.