Pessimism is Unwarranted: Gigabit Networks are Coming

Blair Levin, currently a fellow at the Aspen Institute, is a thoughtful analyst, and wrote the 2010 National Broadband Plan, which was preceded by a thorough and thoughtful analysis of the actual cost to upgrade U.S. access networks for higher speeds.

Also Executive Director of Gig.U, Levin tries to practice what he preaches.

So I was struck by the irony of two adjacent statements in a recent Levin article, the first that the authors of the U.S. broadband plan concluded that "current market forces were unlikely to produce gigabit networks within any reasonable future."

At that time I agreed with the assessment. But Levin says in the very next sentence that "that analysis, and our subsequent discussions with various parties, helped pave the way for the Google Fiber project, which itself has spawned other gigabit projects."

In other words, paradoxically, gigabit networks both were unlikely to develop, and then did develop. Levin believes the plan, and other evangelism, has made a difference. Some of us would agree.

But that perhaps is where practitioners and analysts alike have been too pessimistic, when history might have suggested optimism was warranted.

To be sure, human beings and organizations must act. Human agency is vital. But history itself, in face of what might have appeared to be stubborn agency resistance, would have suggested that, in the face of all obstacles, huge progress would be made.

For all of the apparently sluggish movement towards higher speeds, even in the U.S. market, where complaints abound, the history of speed increases shows a minimum doubling about every five years, and an order of magnitude increase in speeds (up to two orders or magnitude) about every decade.

That suggests gigabit speeds by 2020 would simply  be in keeping with past speed increases. Sounds crazy. I wouldn't have believed it myself. But the data on the trend is there.

The math is clear enough: when you double the quantity of anything often enough, fantastically large numbers result very quickly.

Levin's sentiments are understandable enough. He knows the cost of networks and the practical chores associated with getting them funded and built. But even Levin acknowledges that what might have seemed impossible suddenly erupted.

I think that is the point. Somehow, against the odds, human agency keeps bumping us along an access speed adoption curve that seems "impossible."

Observers long have noted that Moore's Law does not apply to many physical processes and products. Paradoxically, even in the "you have to dig trenches and hang cable," physical world of construction, it seems Moore's Law is a fairly good predictor of access speed improvements.

It is unexpected. Logic suggest it should not happen. And yet, despite all logic, Moore's Law has been a rather good predictor of access speed improvements.