Video challenges the internet industry

Last week's Webinar on broadband's potential to deliver the HD experience, neatly captured perhaps the most fundamental near term challenge facing the IPTV and online video industry. Colin Dixon from the Diffusion Group summarized the underlying market problem when he told listeners the market is rapidly moving from an expectation of short, low-quality video, to a near insatiable demand for longer-form, near-HD content.

Dr. Larry Roberts--who in the 1960s actually helped construct the first Internet--told the webinar the Internet had never been set up to carry video, nor to discriminate between various traffic flows. The problem, he explained, was that there was no system for distinguishing between time-sensitive data flows such as live streaming, and non-time-sensitive data such as file downloads. This meant every service request was contesting for the same bandwidth, a problem made more difficult by the use of P2P networks, which literally sucked up any available bandwidth. As Roberts noted, this has led to the perverse situation where five percent of users are now taking 80 percent of the bandwidth.

Various solutions were proffered: Robert's company, Anagran, is promoting flow routers to discriminate the traffic, while others argue that smart switching already is solving the problem. Deep packet inspection techniques holds some promise, but many believe the most sensible route is a market solution where consumers pay a premium for guaranteed hi-definition quality. In essence that is where Verizon's FiOS service has gone.

Whatever the solution, the startling number is that video is predicted [1] to drive total U.S. residential traffic demand from around 5 exabits in 2006 to around 80 exabits five years later. Much of this bandwidth flood will be met from dedicated IPTV platforms. But by 2011, streaming and P2P video is expected to account for nearly three times today's entire Internet traffic.

In too many regions of the U.S., bandwidth is already so sluggish to be virtually unusable for video--especially in the evenings when P2P networks are most active. While carriers have ambitious bandwidth expansion plans, the sheer size of the predicted demand for video means the pipes will be filled as fast as they are expanded. Unilaterally managing traffic is an obvious strategy, but as Comcast recently discovered this can create a whole set of other problems.

As the year closes insatiable video demand remains the biggest single problem for the 40-year-old Internet. That--and a U.S. economy heading into sharp decline--threatens to prematurely end the web 2.0 era and all its assorted creativity. For those who missed this most important webinar, log on here [2] to listen as our panelists debate these issues and more. --Tom  [3]