(continued from previous post)

It's a mistake to confuse "speed" of connection with availability of bandwidth. Internet packets comprising a channel such as that required for video streaming travel from source to destination at a narrow range of speeds, generally spread across a number of parallel (and some not-so parallel) paths. More video destinations means that more paths and equipment are required for delivering these streams. This connection equipment, including switches, routers, buffers, converters and electrical, radio and optical conductors and so forth COST MONEY. They cost money to deploy, they cost money to maintain, they cost money to power and cool and they cost money to augment and replace when their capacity is reached.

The populist narrative often repeated that connection fabric providers who aren't required to provide endlessly increasing bandwidth on oversubscribed network segments whose once overcapacity is now overburdened to perky, spunky, "disruptive" little "start-ups" will strangle the next Skype or Netflix or Vonage in favor of the "MAN" is a stupid, childish, feckless canard.

If the public demands that network fabric operators be classified as Common Carriers, such a decision would not an unreasonable one, considering the differences they have and similarities they share with other such services. Categorizing network fabric in this way will have other ramifications that may be seen as positive or negative, depending on one's interests.

But requiring that all source-destination entities (such as those comprised of Netflix and its subscribers) be able to reserve limitless connection hardware at the same price schedule as those who send email or the occasional still image over the same connection fabric is like a "start-up" courier service demanding that airports charge tourists a fee to build new terminals because DHL and FedEx and UPS already lease existing facilities.

This segment was a perfect combination of a journalist innocent of the technological issues involved and duplicitous interviewees interested in confusing the listener.

Since the pre-IP days of computer networking, institutions have paid a lot of money for so-called "leased lines" that exist along side public switched telephone and data connections, for data transmission using SNA and DECnet and similar protocols as well as ISDN and T1 PBX connections. They paid for their high use of these connections and for the guaranty that these connections would be available on demand, even when the lines were not continuously in use.

After the late 1970s, increasing network demand and the introduction of the internet protocol coincident with the breakup of the Bell System in the early 1980s lead to a variety of interests spending money to acquire existing network infrastructure and to begin installation and deployment of ever increasing connection fabric. By the early 2000s, substantial capacity had been established for data connection in the United States. The system included a great deal of overcapacity, to allow for future growth of use.

In the last fifteen years, capacity has continued to grow. However, the development and acceptance of video-over-internet has grown even faster. The share of internet connection fabric that large video providers commandeer has become very substantial very quickly. For a period, the overcapacity built into the fabric could readily accommodate this rapid increase but the available margin is shrinking quickly. Those who wish to maintain their business plans that count on the availability of this infrastructure (and wish to grow further) have realized that they need to become partners with their fabric providers rather than their antagonists. That's why Netflix negotiates with Comcast rather than taking them to court (though they will undoubtedly do that too).

(continued)