Cisco is back with their latest report about the “coming exaflood” set to alarmist headlines in the press.

In the spring, the prevailing theory of one “research group” was that bottlenecks would ruin the net’s usefulness by 2011.  That was the one adopted by Time Warner Cable’s unsuccessful efforts to convince residents in four cities that Internet Overcharging was a good idea.  Last month, Australian breakfast television viewers were dropping muffins back on their plates when they were told the Internet was going to be subjected to a massive traffic jam by 2012.  The date of the potential online apocalypse has been pushed forward to 2013 this month, the last year Cisco covers in their data model.

Of course, all such “exafloods” can be mitigated to some degree by purchasing Cisco products and services to handle the tsunami of traffic.

Companies that have a vested interest in doing such studies, in this case to help spur upgrades, always casts suspicion over the results.

The results of those studies are often sold to advocacy organizations (if not quietly funded by them outright) to integrate into lobbying campaigns.  In the push for “exaflood” panic, some of the lobbying groups seek government investment in broadband infrastructure on behalf of their clients, others want to use the Internet growth argument to prove there is a need to engage in Internet Overcharging to finance construction of improved networks (even at a time when some of those companies enjoy billions in profits and have systematically reduced investment in maintaining and expanding those networks).  Cisco’s interests may be closer to home — generating revenue for themselves.

One man who doesn’t have anything to gain from the results is Andrew M. Odlyzko, who runs Minnesota Internet Traffic Studies at the University of Minnesota, an ongoing project to soberly analyze Internet growth.  Unlike others who have repeatedly warned about Internet brownouts, crashes, and slowdowns, Odlyzko doesn’t have a “dog in this fight.”  Once you strip away the self-interests many others have in promoting an “exaflood” agenda, the simple fact remains: with growth in demand also comes growth in new technology and capacity to meet it.  Odlyzko continues to point towards slowing growth.

“In spite of continuing stories about a flood of video overwhelming the Internet, global wireline traffic shows no sign of moving up from its approximately 50 to 60% per year growth rate. If anything, the trend lines point down, not up,” according to the results posted on his website.  Cisco had to echo Odlyzko’s predictions during this past year, but the company blamed the global economic downturn in their report for the decline in the growth curve.

The Economist also debunks the panic attacks:

Talk of exafloods is nothing less than scaremongering and has no bearing on reality, even though video traffic is increasing substantially, says Grant van Rooyen of Level 3, a company based in Broomfield, Colorado. It operates network backbones that carry around a quarter of the world’s internet traffic. “We estimate that 50-60% of traffic today is video, but it’s been that way for the last three to four years,” he says. “We really don’t think we’re going to see a massive failing of the infrastructure.”

Level 3 has been regularly upgrading its capacity, and will continue to do so, says Mr van Rooyen. “This isn’t like building a toll-road with an inflexible infrastructure,” he says. “In the network world, we are able to scale infrastructure and capacity in real time.” When bunches of optical fibres are laid in the ground or on the seabed, for example, not all of them are immediately used, or “lit”. So the capacity of a link can be increased by lighting more fibres. Even when all the fibres are lit, capacity can be further increased by upgrading the equipment at each end of the fibre. Technological progress means the amount of information that can be squeezed down each fibre is steadily increasing.

Back in 1995 Bob Metcalfe, an internet guru and the founder of 3Com, a network-equipment maker, predicted in a magazine article that the internet would suffer “gigalapses” and grind to a halt by the end of 1996. He promised to eat his words if it did not. His gloomy prediction was proved wrong, and in 1997 he duly put the offending article in a blender with some water at an industry conference, and ate the resulting pulp with a spoon.