The internet is at risk of being overwhelmed by bandwidth-hungry traffic. At least that's the conclusion of a new study from Nemertes Research, which calls on U.S. ISPs to increase their broadband infrastructure spending over the next few years by tens of billions of dollars to prevent a severe decline in service. Does the firm's analysis pass the sniff test?
As enterprises look to migrate their private WAN services to the Internet, the integrity of that infrastructure is of paramount concern. Nemertes, which counts both Fortune 2000 companies and carriers among its clients, exhorts U.S. Internet providers to spend $42 billion to $55 billion more on their infrastructures--that is, 60% to 70% beyond the $72 billion they now plan to spend--or risk serious Internet "brownouts" and application slowdowns. Furthermore, the firm warns that this pending traffic jam will "slow down the pace of innovation" and likely discourage "the next Amazon, Google, or YouTube."
While the Internet's core fiber and switching/routing "will scale nicely to support virtually any conceivable user demand," the firm asserts, "Internet access infrastructure, specifically in North America, will likely cease to be adequate for supporting demand within the next three to five years."
Bold assertions, especially as they're based on a body of research that even Nemertes concedes is murky. The Internet "is almost opaque to serious researchers ... for the simple reason that carriers and content providers refuse to reveal their inner workings," the study states. So how can the firm be so confident in its investment recommendations?
Nemertes president Johna Till Johnson says the "order of magnitude" of the firm's spending recommendations is what's important, as "the difference between $42 billion and $55 billion gets lost in the noise." The firm says its study is "landmark" in that separate teams modeled demand and capacity independently, letting them "decouple the impact of capacity on demand." The firm says it consulted and gathered data from scores of researchers, technologists, futurists, enterprise users, service providers, equipment vendors, venture capitalists, and financial firms about Internet traffic patterns, revenue, and investment plans.
Furthermore, Nemertes says its study is independent, funded by the firm's diverse client base of enterprises, vendors, service providers, and not-for-profit organizations. Yet the two chairmen of one of those not-for-profits, the Internet Innovation Alliance, which bought "distribution rights" to the findings, are quoted prominently in the Nemertes news release. A check of the IIA Web site reveals an organization whose members include Alcatel-Lucent, AT&T, Corning, Nortel, and other telecom infrastructure suppliers that would surely benefit from tens of billions of dollars in additional Internet infrastructure investment.
Johnson insists that the IIA had no input into the research and that Nemertes embarked on the study four months ago with no agenda, though she concedes that the firm's news release might give too much weight to the IIA partisans. Still, Nemertes isn't the first to warn that Internet capacity won't keep pace with soaring traffic loads. Way back in 1995, for example, Ethernet co-inventor Bob Metcalfe famously predicted that the Net would collapse within a year--but then more than a year later he proceeded to literally eat his words at a conference by consuming the paper column in which he predicted the demise.
Nemertes emphasizes that it's not predicting an Internet collapse: "Instead, the primary impact of the lack of investment will be to throttle innovation--both the technical innovation that leads to increasingly newer and better applications, and the business innovation that relies on those technical innovations and applications to generate value."
Probably not words Nemertes will ever have to eat, even if after chewing on them for a while we're less than convinced that the situation is so dire.
VP/Editor In Chief
To find out more about Rob Preston, please visit his page.