I'm really beginning to hate the Internet. Back in 1989 it seemed limitless, with incredible amounts of information just waiting to be found and gobbled up. There were never any delays or network outages. Today, the Internet is vast and vastly overloaded. Delays are common and outright blockages occur daily.

Those of you who chose VPNs over the Internet are surely feeling the fire. Branch-office connectivity is becoming ever more fragile, and user satisfaction is approaching all-time lows. On the other hand, those of you who chose VPNs over a single public carrier are wondering what all the fuss is about. Sure, you're "on the Internet," but your traffic is being carried over a single managed, homogenized, monitored network where outages mean lost revenue to the carrier and are to be avoided rigorously.

There's a trend here. Consumers and risk-accepting companies will continue to use the unreliable, insecure Internet as it continues to deteriorate. Risk-averse companies will desert the Internet for better transport models. I predict that within 18 months we'll see an entire day with no domestic Internet connectivity. The Internet will be repaired, but public confidence will be shaken and corporate flight to other methods will accelerate.

As a remedy, large-scale carriers are pushing their own semipublic networks, with all data carried on a common network infrastructure but individual customers' traffic segregated and VPN service available to those willing to pay. These carriers will transport Internet-destined data over their high-grade networks to an "exit point" close to destinations. If the Internet destination is also on the carrier's network, your traffic will never see the Internet. Major players in this market are AT&T Corp., Sprint Corp., MCI Worldcom Inc. and new players such as Deutsche Telekom AG and Broadwing Inc. At the high end, companies with the need to securely push large amounts of information will use the new generation of transport providers, such as Yipes Communications Inc. and XO Communications Inc.

That's the model for the future. The Internet will continue to handle "consumer grade" traffic, but transport will be unreliable and insecure (and priced accordingly); large-scale public/private nets will be offered for reliable, secure access, priced attractively to businesses. The two network types will interconnect, but only at specific, tightly controlled locations.

What does this mean for us management types? If you haven't moved away from the concept of private data networks, you're better off than the rest of us. Those who have embraced using the 'Net for connectivity, with or without VPNs, need to re-evaluate that model based on risk/reward, with risk being user dissatisfaction at balky connections and reward being lower prices. If your evaluation puts you on top of the risk curve, then it's time to think about public/private connectivity or going back to your own private data net. Wherever you fall, it's time to take a good, hard look at your network strategies.

Shapiro is district technology coordinator for Kingsport City Schools in Tennessee. He can be reached at jshapiro@kpt.k12.tn.us.