In a post titled "Internet Traffic Growth Doesn't Matter," Andrew Schmitt from Nyquist Capital raises some tough questions about popular assumptions for network traffic, particularly from Internet video. Andrew is recapping the Gilder TeleCosm conference, and in particular a presentation by Andrew Odlyzko from the University of Minnesota.
Odlyzko will be familiar to many readers with memories of the dot-com boom and bust, as he was among the first to challenge the oft-repeated contention that Internet traffic was "doubling every three months." In his presentation at TeleCosm, Odlyzko observed that "Internet traffic growth rates are slowing. Hype is accelerating."
Some of the hype is intertwined with the network neutrality debate and the assertion by some analysts and providers that Internet capacity will be strained by the robust growth of online video and large-file data transfer. Odlyzko's Minnesota Internet Traffic Studies (MINTS), which aggregates traffic data from a number of sources, estimates overall Internet growth rate at 50 to 60 percent per year. His presentation suggests a mixed picture - there are huge potential bandwidth drivers on the horizon, but Internet traffic is actually slowing in several high-bandwidth markets.
That's the case in Hong Kong, where per-capita Internet usage is about six times that of the U.S. Odlyzko's data shows that the Internet traffic growth rate in Hong Kong has been slowing since 2004, with growth of 11 percent in 2007.
Schmitt says his analysis of broadband traffic in Japan, one of the most mature markets for fiber to the home, shows that per capita bandwidth consumption is growing at an annual rate of about 18 percent. Decent growth, but not the 50 to 85 percent annual growth seen for the U.S. and global Internet.
Given these data nuggets, what assumptions can be made about bandwidth growth in the U.S. as broadband penetration reaches the levels seen in Japan and Hong Kong? Schmitt argues that assumptions need to be either challenged or supported by more data:
The consensus opinion that video traffic driving bandwidth demand is unchallenged, but is remarkably similar to claims in 2000 that Internet traffic was doubling every 90 days. I would really like to see someone objective, like Andrew Odlyzko, look into the claim that video is driving the growth of the network.
Schmitt concludes: "Traffic growth simply doesn't matter. Period. What matters is revenue. Arguments about an 'Exaflood' or Video bandwidth explosion are not meaningful unless their deployment results in incremental revenue that carriers can use to fund new investment."
There will be plenty more debate on these points in months to come. The data outlined by Schmitt and Odlyzko add some interesting reference points.