What is not okay, is if a company charges you differently for the same service. 100 MB of HTTP-surfing cause the same costs for my carrier than 100 MB of FaceTime. The only difference is that FaceTime potentially decreases a different revenue stream in the same company, namely normal voice calls.
No, that's most certainly *not* the only difference. Facetime, according to some estimates, uses approximately 4MB/minute of bandwidth. Typical web-surfing uses significantly less than 4MB/minute of bandwidth, like on the order of 40-100KB/minute. That's 1/100th-1/40th of the data rate for web-surfing vs. Facetime use.
For a reasonable analogy, let's stick with electricity because Data and Electricity both share some common traits:
1) Over a long-enough period of time, availability is essentially limitless for a single user.
2) There are limits to minute-to-minute availability which can be impacted by things outside of the direct control of either the user in question *or* the provider. (That is, other people, by using too much of either all at once can result in not having enough available for our sample user.)
So, we have our wonderful, unlimited plan before the advent of Facetime or the washing machine. (Two things which consume larger-than-typical loads of their respective resources over significant, but not indefinite periods of time.) Everything is good so far, because typical loads are kept fairly even by the nature of the devices using the resources and the law of large numbers.
Now, suddenly, we have Facetime and the washing machine. Both of these gobble up comparatively large amounts of our time-limited resource, drastically increasing the minute-to-minute demand of a single user. If only a few people do this at once, it's not a terribly big deal. If they become popular (like the washing machine did), then suddenly there's the potential for a real problem.
In the case of the washing machine, it's a usage-pattern problem. Back when it was introduced, the hubby went off to work in the morning, and the kids off to school. The wives, would then start their daily chores, and housekeeping. If many washing machines are run at the same time, then suddenly there isn't enough minute-to-minute juice in the lines to keep everything running, and you get brownouts and blackouts caused by the sags and surges involved.
The same thing can happen with data networks. If a high-bandwidth-demand service becomes popular, it has the potential to cause the same issues related to over-demand. Now, bandwidth availability can be improved by upgrading tower back-hauls and adding towers, but there's significant lead-time (often more than 6 years) involved in getting a new tower approved, and built. Most of that time has to do with finding good sites, negotiating leases or purchases, and getting the required permits. Once all of that is done, the actual install can take place over the course of a month or two. Little of the delay involved in improving the networks is attributable to the network operators. Instead, most of the delay comes from dealing with bureaucratic red tape.
To combat this problem in the US (and to aid with getting as much area covered as quickly as possible in the beginning), the network build outs used as few towers as possible to get the job done. This is different than in most of Europe, where they built towers more densely.
Now, the US networks are in a tight spot. Overall demand has recently increased by better than 2 orders of magnitude, and some available services go another 2 orders of magnitude above that. To combat this they're upgrading their existing infrastructure as quickly as possible, but they also need to add towers. Unfortunately, adding towers takes *years*, instead of months, so this process is slow.
Remember the existing physical network infrastructure was largely designed and built around *just* voice transmission, using technologies chosen because they covered as much range as possible with as few towers as possible. Most of the newer voice/data protocols use frequencies which *don't* travel so far at the same power level. This means new towers have to be built. And those towers *are* being built, but the network providers actually don't have the final say for when and where towers can be added, so it's going slowly.
I'm lucky. I live in an area with decent tower coverage for our population density. There's a few stray dead- or weak-spots, but by and large things work well, and they've been able to keep up with overall demand pretty nicely. My sister lives farther from a major city, and only recently got 3G service *at all* because new towers had to be built to provide it, and it took 3 years longer than the network provider's initial estimates to get all of the red tape cleared away so they could build the dang things.