Prompted by a reporter's post on Dave Farber's listserve and Jim's question about how net neutrality will affect the 'computing-as-a-utility' model, I offer this slightly jaded appraisal of how net neutrality regulation would end up playing out in a world where both public and private networks co-exist.

First, there will be taxonomical distinction made between public and private networks. Part of a brodband pipe will be treated as "public" and regulated as such, while another part (for instance, the part carrying video programming) will be treated as private and not subject to the panoply of "public Internet" net neutrality mandates. While there will be no principled line that can be drawn to make this distinction, it will be necessary to avoid the more draconian effects of net neutrality.

Gradually but inexorably, everyone with specialized needs for QoS, low-latency or the like, will be purchasing a "private" network capacity that is not subject to the public Internet net neutrality rules. Microsoft and Sony, for instance, might buy or subsidize "private" broadband capacity so gamers on their network will get the low-latency service they need for interactive gaming. Net neutralists will scream at this legerdemain, but will be powerless to stop it because, at some point, even the courts and regulators will have to recognize the validity of private contracts for private service.

There is historical precedent for just such regulatory bracket creep taking place. Both "private line" service

with telephony and "interruptible service" for electric utlities amount to categories where the regulators save themselves from their own mandates. With private line service in telephony, regulators were compelled to invent this category so businesses' communications needs could be customized and met without the onerous regulation and cross-subsidization inherent in the PSTN. In truth, of course, a private line looked no different than a public line — the same copper could carry both private and public lines. However, you needed to have these private, relatively unregulated arrangements, or the customers would bolt the network. By leaving the network, all consumers would be worse off, so 'voila' all of a sudden "private line" service showed up absent the traditional regulatory burdens. The same is true of interruptible electric service to larger users. In truth, interruption almost never happens, but it allows price discrimination in favor of heavy users that might otherwise leave the grid.

In the end, then, a good hardened cynic sees net neutrality as redounding to the comminications bar's benefit -- with all the associated transaction costs -- but to no one else's. The Internet will be differentiated with different tiers of service and two-sided markets will develop as more effective price differentiation reaches different consumers. This will be done by gradual differentiation between "public" and "private" offerings, where all of the commercial freedom lies on the private side of the pipe. Of course, regulators of will could block such an outcome, but the negative investment effects and the needs for low latency, high quality of service would drive them to allow such arrangements.