Can Carriers Afford "Unlimited" Internet Access?

With the caveat that retail price is not necessarily directly related to cost, it sometimes might be possible for an Internet service provider to offer truly limitless Internet access, for a while. That tends to happen only when a brand-new network, operated by an upstart trying to take share, first comes to market.

Under those conditions, the network has plenty of capacity, and it is customers that are in short supply. Assuming that service provider actually is able to make a business of Internet access, the number of customers soon will force a change in such policies at some point, to encourage customers to make choices about how much they use.

Also with the caveat that some think there really is not a bandwidth issue faced by major ISPs, Internet access actually is more akin to many other network services, and unlike a few that charge the equivalent of "flat rates."

The classic example of a "flat rate" network service is cable TV, which allows users "unlimited" use of all the channels, on multiple sets or devices, for the same charge. But that is a multicast service, which has important bandwidth implications. 

Essentially a multicast service sends one copy, usable by all users, on a point to multipoint basis. Satellite TV, broadcast TV and broadcast radio work the same way. Metering doesn't matter, because there are zero bandwidth implications for usage, high or low. 

No more bandwidth actually is consumed no matter how many customers decide to watch at once, or how many customer devices are in use at any one time. 

Internet access is a point-to-point medium, like a voice call, a videoconference, water, waste water or electrical service. There, each unit of additional usage really does have network implications. 

At least in part, that is why unlimited Internet access is disappearing. On networks with high usage, there are "peak" hour and "peak day" dimensions to usage. And though it often seems "unfair" to some observers that usage is metered at some level even when the networks are lightly used, that isn't the point.

Every communications network (point to point) has to be sized for "peak" usage, not "average" usage.

So ISPs will have to shift to some form of metered usage over time, even if only to encourage people to use the network at off peak hours, rather than peak hours. Mobile ISPs have been bigger problems, as some parts of their networks have very-high usage on a sustained basis, while other parts are more lightly used.

Any cell site right beside a major highway will become congested during rush hour. Some suburban cells might not become seriously congested at any hour of the day. Some downtown urban sites with high foot traffic might be heavily used during working hours, but lightly used during non-work hours, and on weekends. 

The whole point of metering is that metering also allows service providers to create incentives for using the network at times when capacity is not a problem. That's the whole theory behind weekend and evening calling  rates, which were important in the past. 

The other problem is video, which has bandwidth implications an order of magnitude or two orders of magnitude greater than voice, for example. While nobody seems to think retail rates actually can rise by an order of magnitude or two orders of magnitude, even if usage does grow that much, the additional usage carries real costs, and those costs have to be recovered. 

So can a typical ISP, with a serious number of customers, actually afford to offer truly unlimited access, if those customers start watching Internet-delivered video? Probably not. An order of magnitude worth of network is an expensive thing. 

One might hope more-efficient suppliers might enter any local market, but any large ISP will over time, experience higher costs "per unit," even if that supplier originally started out as a low-cost supplier. 

We can quibble about the cost elements, but those elements will develop, over time, especially in highly-competitive markets such as Internet access.