Reducing The True Cost of Broadband Connectivity

In common business parlance the cost of a product is the expense incurred in producing it. This is distinct from the price paid for it by the consumer. By this definition, the cost of broadband Internet connectivity is high. The impact of subsidies such as those provided by the federal Affordable Connectivity Program (ACP) is to bring down the price of broadband to some consumers --- it has no impact on the cost of producing it.

The ACP has no termination date. The total funding available for the program on Jan 1, 2022 amounts to $16.4B, which will last no more than 2 years. If ?these funds are exhausted the subsidy will stop unless Congress appropriates more money for it (see Doug Dawson's Pots and Pans blog How Long With The ACP Last?)

In a legislative environment some Republicans are proposing to “sunset” all programs, including Medicare, after 5 years. We cannot take the continued subsidy of broadband connectivity for granted. It is all but certain that the ACP will eventually be cut back, its means testing perhaps made stricter. Sooner or later, it will disappear into the metaphorical sunset that increasingly awaits all economic support for low-income Americans. At that point the price of broadband to all Americans will have to reflect its true cost.

Why is the true cost of broadband so high, and can anything be done about it? The answer is more complicated than it may seem at first. All Internet applications are carried over the same wires employing the same fundamental transmission protocol, but there are different ways of using these resources. “Transfer applications” move files (sometimes called digital objects) across the network and then store and process them at the receiving end. Examples of transfer applications include browsing Web pages and downloading programs or videos. By contrast “interactive applications” use the network to connect to a remote server with which it exchanges information on a continuous basis. Examples of interactive applications include teleconferencing and remote gaming.

What is not widely understood is that a “transfer network” that can support only transfer applications can be built and operated much more cheaply than an “interactive network” that also supports interactive applications. The true cost of such a network can be much lower, and so the price can also be lower even without high levels of subsidy. Transfer applications can provide very high levels of service, including high bandwidth delivery of streamed data. But certain popular interactive applications require an interactive network, whose true cost is much higher.

To be sure, interactive applications seem a lot more excitig than transfer applications. How valuable can transfer applications be? This question is hard to answer, because some applications that could be implemented as transfer applications are currently implemented interactively. There are programming techniques called “latency hiding” that can be used to implement an application using a cheap transfer network to allow it to provide a user experience which is the same as, or close to, that of using an expensive interactive one. Today latency-hidig techniques are often not used because it is easier to rely on interactivity, and the cost of the interactive service is being paid by the consumer or the taxpayer. However most applications other than highly interactive ones like teleconferencing and remote gaming can be implemented using such latency-hiding techniques.

But how low can the cost of a transfer network be? Again, this calls for such speculation, but there is reason to believe that the cost might be low enough that access to non-commercial transfer applications could be could be free. The cost of a transfer network might be low enough to allow it to be supported by a fee paid by the providers of for-profit applications without overly burdening online commerce. In such a model the true cost of providing access to critical applications required for digital citizenship could be affordable or free permanently.

What would the impact be on the price of interactive connectivity? If end users chose to use only low-cost transfer networking, instead of paying a high (or subsidized) price for interactive connectivity, there would be fewer paying customers supporting the cost of the latter.?In effect end users who do not require interactive connectivity are currently paying (or receiving a subsidy) to keep its price lower. On the other hand end users who currently struggle (or require subsidies) to afford Internet connectivity might get free access to critical digital services needed for participation in a modern connected society.

For further discussion see this recent article: Is Universal Broadband Service Impossible?, Micah Beck and Terry Moore, arXiv:2204.11300, April 2022.