Network Computingwww.networkcomputing.com

Beyond the real-dollar outlay for WAN links, there is enormous waste in our scramble to reduce service costs. We rip out perfectly good equipment to slice ongoing expenses by a few percent. Shifts in carrier tariffs create and destroy entire product lines. Developers are forced to build, and rebuild, applications to respect low-bandwidth links. Only recently have top-end client/server applications achieved decent performance on a WAN. Network administrators adopt elaborate schemes to analyze and distribute costs back to the user. This is complete insanity, and nobody's running the asylum.

Meanwhile, the cost of providing WAN services has plummeted. DWDM (Dense Wave-Division Multiplexing) equipment lets carriers run 100-Gbps traffic on the same fiber pair that carried 2.5 Gbps a year ago. HDSL (High-bit-rate Digital Subscriber Line) technology has knocked the cost of installing T1 services through the floor. The per-port cost for switching equipment has plummeted. And standard platforms have reduced the cost of managing networks, or at least held it constant.

Nonetheless, the price of a T1 hasn't dropped substantially for the customer. In fact, it recently has begun to rise. Telecommunications competition has become just another legal dance, and there is no relationship between the cost of providing a service and the ultimate cost to the customer. How much is a T1 worth? How much would you like to pay?

Granted, there is more to telecommunications services than simple bandwidth. There are many intangibles to consider, including reliability, availability and responsiveness; and more tangible areas such as billing and customer network management. The value curve runs from cheap and minimal to expensive and elaborate, and each of us must consider what best fits our needs.

Intangible Benefits? At the top of the value curve you'll find the expensive offerings from the major carriers; at this level, the intangibles dominate. Carrier representatives blather on about "leveraging technology via strategic business partnerships in the competitive global market" and other impressive-sounding strategies. Half the time you don't have a clue what they're talking about, and don't be surprised if they're not sure themselves.

ISPs dominate the lower end of the value curve. Many simply sell pipes, playing on the accepted gap between volume wholesale and single-line retail pricing. Some sell other services, but availability, reliability and consistency aren't what you'd expect at the top end. By the way, the top-tier ISPs don't belong here; elite ISPs are simply major carriers with a different technological bias.

Trust drives carrier selection much more than cost--especially at the top end of the value curve. When their business will grind to a halt without its network, you can bet customers will spend money--and lots of it--to ensure that the network stays up and running. But don't kid yourself: Every carrier, and every customer, has had a network outage caused by a WAN failure. When outages happen at the low end of the curve, nobody's too surprised. If AOL goes down for awhile, David Letterman gets some new material. But outages at the high end breed catastrophe. Just this year we saw a major failure in the AT&T Frame Relay network and multiple outages created by PanAmSat's Galaxy 4 satellite breakdown.