ISPs Struggle to Deliver Video Traffic While Imposing Bandwidth Caps

LONDON, May 3, 2010 – Many Internet service providers are struggling to cope with the rapid growth of video traffic while trying to put a reasonable limit on the amount of bandwidth consumers can use.

Bandwidth capping is seen as a way to help pay for new infrastructure by constraining demand on the network while extracting revenue from higher net worth customers who download the most data. However, online applications tend to evolve to fill up the capacity available, creating a situation where there likely never will be a point when pressure on bandwidth within the network ceases.

But bandwidth capping also has a quality of service function, to ensure there is enough capacity for critical applications that generate less data streaming than video, such as e-mail.

“This phenomenon of fast increasing amounts of video over broadband is going to cause some very significant problems in terms of cost and also user expectations over the next few years,” said Simon Orme, director of content services at British Telecom’s wholesale division. “We see these huge hockey sticks from the analysts projecting explosive growth, which get worse every time I look at them.”

When Comcast introduced its generous capping limit of 250 gigabytes in 2008, it highlighted that this was enough bandwidth for 50 million e-mails a month, 62,500 songs or 125 standard definition films a month. That sounds splendid, but that’s only enough capacity to watch about 60 hours per week of high-definition television. That’s less than the amount consumed per week now by the average U.S. household – even before allowing for multi-screen viewing.

Some commentators greeted the introduction of Comcast’s bandwidth cap as the end of the Internet as we know it by expressing concern that other firms would set more draconian caps.

The reality is that within the decade, the Internet will become the vehicle for distribution of all digital content, including the video and TV services currently still delivered within the walled garden of proprietary distribution networks, mostly satellite and cable

The physical network may still be cable or satellite, but it will be an IP-based infrastructure, with the content arriving “over the top” rather than within a walled garden. Access to the service will continue to be controlled. However, content providers now will be in direct contact with the end customer, in effect cutting out the broadcast distributor.

Current TV operators will either morph into Internet service providers, which many are already anyway, or into content providers in their own right. This trend has been happening for a long time in the case of terrestrial services with the separation between broadcasting and transmission, but cable and satellite operators are still often largely distributors rather than content producers, and smaller firms risk being squeezed out of the distribution chain.

This is why the current network neutrality debate is so important. It is partly a battle for long-term control over content and its distribution. This mirrors the likely trend in broadband infrastructure, which will connect content producers directly with the consumer, or at any rate with the access network. This is the likely consequence of the bandwidth crunch from which both fixed and mobile operators are currently suffering, with the squeeze increasingly tightening on the backhaul and core networks.

Some telecommunication firms have been quicker than others to respond to the challenge through a combination of upgrading the physical infrastructure to generate more capacity, and creating a new overlay for video content in particular.

BT in the United Kingdom for example is building a content distribution network called Content Connect in a joint venture with two unnamed ISPs, and several broadcasters including the BBC. This was an urgent response to the surge in video traffic generated by catch up services such as BBC iPlayer clogging up core Internet capacity around the world.

The answer is to create dedicated video delivery networks overlaying the existing infrastructure, ultimately feeding directly into the access network. Orme admitted this in turn also would prove disruptive for businesses by changing the current relationship between digital content production, distribution and consumption.

But this may not quite mean the end of the current model for TV distribution, according to Ronni Zehavi, CEO and co-founder of global content delivery service provider Cotendo. Zehavi argued there still will be a need for specific content delivery solutions tailored to different industries and even major content producers.

There will be significant variations in the distribution profile of sources and points of consumption around the world, as well as in the form of the content itself. The nature of the web acceleration required will differ significantly, given that some content will still be for local consumption, some regional or national, and some global.

The likely outcome could be a split between underlying physical infrastructure, which may not distinguish between the data transmitted over it, and the network level routing or distribution. But one thing is for sure, video and TV are moving to the Internet, and various forms of  content distribution networks will evolve to cope with the resulting deluge of sustained traffic.