Optimal Pricing for Integrated Services Networks
The concept of an "integrated service network" describes a single network that provides different services (data, voice, video, etc.) through a common set of wires, enabling suppliers to leverage economies of scope.
However, heterogeneous services have significantly different levels of demand, regarding their bandwidth consumption: a person watching high-definition television requires much more bandwidth than a person who uses it for intermittent voice communications; and the former requires a mode that receives far more data than it sends, whereas the latter communicates roughly equal amounts of data in both directions.
The question the author posits is how the pricing of these different services should differ, given their disparate impact upon the global data network, with a goal of a pricing structure that gives priority to the kinds of traffic that "maximizes social welfare."
EN: same problem as the previous five chapters, though the goal of his solution seems a little odd. Perhaps this is a socialist perspective?
The Network Service Model
Fundamentally, the "network" is a configuration of point-to-point connections over which traffic for various services is routed. Each of these services comes with some "guarantee" of quality, and fulfilling that guarantee requires different levels of service. In addition to the differences in of volume of data, duration and frequency of transmission, there aer two other factors:
- The loss-tolerance of the technology (if a frame or a few milliseconds of audio is lost from a video, the video is still watchable, and the user may not appreciate the loss in quality, but if data is lost from an executable program, the entire program will fail).
- The time-sensitivity of the technology (if some data does not transmit at the beginning of downloading a program, it can be downloaded later and added at the correct place; but if a word is dropped from a telephone conversation, it cannot be transmitted later to "patch" a gap that occurred earlier)
Unless network capacity is infinite, there will be instances in which capacity is insufficient to serve all users' needs for all the types of media. Accounting for the factors above, how does one prioritize traffic, by implementing criteria that define which traffic is to be given greater priority?
The Optimal Pricing Policy
The optimal pricing policy is mathematically described thus:
I'm not going to explain that in detail, but it's a method of developing a weighted-average cost that considers not only the amount of data transferred, but also how that volume compares to all other traffic at a given instance, considering the types of transmissions.
By implementing such a system, a supplier would offer pricing on services that reflects their bandwidth requirements of specific media. Unless other suppliers did likewise, this provider would attract users of low-bandwidth media, who would be offered a discounted price. This would lead the provider to increase the unit price of that service to cover their (fixed) network costs, until supply and demand agreed on an equilibrium price.
Meanwhile, a supplier who fails to implement a pricing schema that fails to consider the bandwidth requirements of specific media will offer a lower price than others on high-consumption media, attracting users of that media until his system capacity fails to serve their needs, at which point he would be compelled to raise prices, with a goal of either raising the revenue needed to extend his network capacity, or shed enough customers that his capacity would be sufficient for those remaining.
The net result would be equilibrium pricing for all services, regardless of whether each network provider structures his own prices to accommodate various kinds of traffic, or whether certain kinds of customer are attracted to specific providers in order to obtain fair pricing for their service (in which case, providers would specialize in a specific type of network traffic).