It is going to end the Internet as we know it. It will save the carriers and let them build the Internet we need. 'Net neutrality is a big issue, with everyone from network experts to politicians weighing in.

'Net neutrality basically means that the carriers treat all traffic equally. Carriers are questioning this logic and want to be able to charge extra for better treatment. Sites that pay more would get a higher priority; those that don't would go to the end of the line.

The problem with 'Net neutrality arguments is that they miss an important point: Can the carriers provide differentiated service? They can set priorities, but does that mean people will see the difference - that it will really matter? I don't think so.

Carriers would implement 'Net non-neutrality through QoS and bandwidth management. Both have little effect except during high use. When use is at 62%, there is an average of one packet in the queue; even at 75% use the average climbs only to a few more than two packets.

Assume a packet arrives at a bad time, when the queues are above average length. At 60% use, there will be only about two messages in the queue 95% of the time; at 75% use, the queue would lengthen to about eight packets. That may sound bad, but it really isn't. Assume a T-3 line is used - not very fast by today's standards - and all messages are 1,500 bytes - near Ethernet's maximum. The wait adds only a few milliseconds.

How often are lines at 60% to 75% use? Not very often. Most carriers have engineered their networks to run at lower use, and it is high only for short times during peak hours, if at all. The majority of the time, networks are running at lower rates, with no or very small queues.

The result is that paying for premium service over the Internet backbone would be paying for nothing. You would get the same level of service at the inexpensive rate. But is there a way for the carriers to make it have a difference? Yes: They could cause problems artificially. For example, they could make the allowable queue length for low-priority traffic very small, causing lower-priority messages to be discarded at an unnaturally high rate.

Even this might not work. Any carrier that does this would affect a lot of traffic, because most users would not pay the premium price. The word would get out these carriers run a poor-quality network. Enterprises, DSL and cable providers would hesitate to use them.