A lot of folks have reacted to Chris Anderson’s deliberately provocative piece in Wired: “The Web Is Dead, Long Live the Internet.” I have two chief reactions. One is a methodological one — Anderson gives no justification for reliance on percentage of total Internet traffic as being a measure of anything in particular from which we might draw conclusions. I am hardly the first to note, for example, that according to Anderson’s chart, DNS traffic ceased to matter by the mid-1990s, a conclusion dramatically contradicted by actual reality.
But my chief criticism is substantive. Anderson — perhaps unintentionally — does an excellent job recapitulating Karl Marx’s original Socialist critique of capitalism, i.e., that it will invariably reduce to a monopoly or cartel structure exacting monopoly rents (although he leaves off the part about it eventually collapsing under its own inefficiency, the workers seizing the means of production, yadda yadda yadda). But his conclusion is that such is human nature and we ought to just suck it up as long as we keep getting cool stuff. (Aps are the opiates of the technorati masses, apparently).
But there is a reason I am not a socialist (despite claims of some critics to the contrary) and instead brand myself as a member of the Congregation of the Progressive Capitalists. Anderson notes that “Monopolies are actually even more likely in highly networked markets like the online world. The dark side of network effects is that rich nodes get richer.” But he overlooks the ability of public policy to prevent that from happening. Anderson appears ignorant of the role of such things as the FCC’s Carterfone decision and subsequent rulemaking, or the role of the Computer Inquiries in creating the conditions for the growth and development of the Internet and the applications that ride on it, including the Web.
Accordingly, if we ignore the methodological problems and accept the underlying economic argument, the solution is not to develop ill-suited analogies based on the happenstance that we can somehow define “the Internet” as “post-adolescent” to somehow rationalize our loss of freedom. To the contrary, if we are really seeing the decline of the Web and the rise of the App, we have a policy choice to make. We can do nothing, and follow Anderson’s inevitable slide from the open world of the Web to the closed world of the Ap. Or we can do what we did to the wireline world 40 years ago in the FCC’s Carterfone and Computer proceedings and wedge the system open.
Put another way, we can still save the vibrant free market on the web through a little proactive regulation, rather than accept Anderson’s “inevitable” collapse.
Let’s start with the methodological criticism. According to the chart, DNS traffic stopped in the mid-1990s. When last I checked, we still relied on the DNS.
What Does Percentage of Total Traffic Really Tell Us? Not Much.
This highlights a number of problems with using the relative level of traffic as a measure. Traffic may surge or drop relative to other uses for a variety of reasons. For example, the percentage of traffic representing DNS lookup and routing dropped dramatically relative to all other traffic because DNS got more efficient, caching became more common, and other forms of traffic rose exponentially. But if we had concluded that this reflected the decline of the importance of DNS (as some actually did in the mid and late 1990s), and postulated that the rise of search engines made domain names obsolete, we would have been utterly wrong. At the time when DNS traffic was “declining,” registration of names was booming, and the DNS-based web was growing exponentially.
Similarly, relative volume of traffic doesn’t tell us about the relative importance of the traffic. For example, if what apps do is collect information from websites, that is interesting — but all you’ve really done is introduce the equivalent of a new browser. None of this tells us what _changes_ in a real sense.
Is Monopoly Control “Inevitable” As Anderson Suggests?
Which brings us to the real meat of the article — that the change in relative traffic somehow represents a “maturing” of the Internet. Granted, I start skeptical. As someone active online since 1984, and active in internet law and policy for nearly 15 years, I have lost count of the number of times the Internet has supposedly “matured” — where “mature” is a code word for “things will be less good than they were.” It goes in the same box of cliches as the Internet no longer being a “wild west.” Mind you, the same folks who push the line that sadly the wonderful world of an open internet must “mature” and “grow up” by doing the equivalent of putting on a suit and tie and going corporate will reverse on a dime and extol the wild, open innovative net that is still nascent and dynamic and must at all cost be protected from any government regulation. But I digress.
The more substantive criticism of this “insight” is that it falls into the basic fallacy of assuming that the forces shaping the mobile world — regulation, economics, and technology — are somehow indifferent natural forces that must yield an inevitable result. This is as much nonsense now as it was in 1998 when ISPs like Digex were telling me Washington was irrelevant because they were the small and cunning mammals who would triumph over the giant stupid dinosaurs like AT&T and Comcast.
The truth is that all of these forces — economic forces, social forces, and regulatory forces — are interrelated and influence one another. The open internet of the dial-up world evolved as it did because we had a bunch of federal rules in place that required the telephone company to allow any attachment to its network, required it to sell services to rival businesses, and prohibited it from messing with any traffic going over its system. We did not impose these rules on the wireless side. As a result, the two world have evolved very differently.
Chris is right that people will cheerfully adopt what works and what is easier. This is how smart businesses make a living and why the app stores are so popular. But people also make decisions based on cost and availability. As long as there are things available on the open “web” Internet you can’t get on the closed mobile app net, the web will continue and apps will continue to chiefly mirror web-based functions — with apps simply serving as specialized browsers that charge for the convenience of mobility and cost somewhat more to use. If policy permits the broadband access providers that are the gateway to the web to replicate the closed model of mobile platform, we can certainly expect to see the web decline as it becomes indistinguishable from the mobile appverse. By contrast, if we had rules that opened the mobile universe, we would see it start to develop in new ways that more closely resembled the “immature” Web.
It’s worth comparing some of the examples of network monopolies Anderson sites and see how policy altered their impact, even if the underlying distribution technology became a “natural monopoly.” Before we had any theory under which federal or state government regulated commerce, we saw the rise of the telegraph monopoly and railroad cartel. What made these particularly devastating from an economic (and civil liberties) perspective was that unregulated control of a critical means of distribution beget additional monopolies in goods and services dependent on that means of control. Thus, for example, Western Union’s monopoly control of the telegraph allowed it to foster the creation of the Associated Press, which in turn became a monopoly on the ability to collect news and share correspondents between newspapers until the United States successfully concluded an antitrust suit against the AP in 1945. Similarly, the railroad cartel enabled Rockefeller to create Standard Oil by providing him with discount shipping so that he could easily undercut the prices charged by potential rivals, allowing Rockefeller to either buy them out or break them.
Now flash forward to the regulated monopolies of telephone and power. True, the distribution system itself reverted to “natural monopoly” as Anderson describes. But the impact of these monopolies was considerably different because the monopolist was required to treat all traffic equally. Neither the monopolist nor its customers could leverage the underlying natural monopoly in the way the Associated Press leveraged Western Union or Rockefeller leveraged the railroad cartels. For example, Sears Roebuck was never able to team with AT&T to give itself an advantage in its mail order catalog business model. Instead, as the telephone emerged as a means of mass communication, rivals could adopt the model and challenge Sears for supremacy. And, alongside Sears and other mail order giants, businesses around the country were able to develop competing businesses utilizing the telephone for things like Chinese take out, delivery services, or any other business model that made sense.
Similarly, while we tend to discount the importance of the Microsoft antitrust case because it ultimately resulted in a slap on the wrist, it limited Microsoft’s ability to leverage its control of the desktop platform at a crucial time in the Web’s development. For one thing, MS felt compelled to keep Apple alive and bring back Steve Jobs as a means of showing the government competition continued to exist on the desktop — a move with pretty far reaching implications for the state of competition today in both wireline and wireless access. But it also curbed Microsoft’s most aggressive tactics against potentially competitors, particularly after 1998. A move that, among other things, allowed the development of rivals such as Google.
(I should note for those who will invariably take the above paragraph out of context that I do not, by any stretch of the imagination, regard antitrust as sufficient to ensure an open Internet. Nor, I think, would Netscape accept the idea that antitrust provided sufficient protection or even compensated them after the fact.)
I don’t think Anderson can prove his case based on traffic stats alone. The Web remains incredibly important, which is why network neutrality on the wireline side remains worth it. Those who will invariably argue that it is “not worth it” to protect an open Internet because the web seems fit to vanish into obscurity are as wrong as those who told me in 1997-98 not to waste my time on
ICANN because the days of DNS were numbered. At the same time, those who consider it “inevitable” that as the Internet “matures” it will devolve into a less open environment because “that’s what people want” are equally mistaken. We have a choice, just as we have always had. If the Internet is supposedly 18, perhaps it will take the example of its 40 year old Carterfone parent to heart. A little openness regulation keeps your arteries open and keeps you young and innovative.
Stay tuned . . . .