Building It Out

Despite the burden of surging traffic and attacks on its very foundations, the internet has yet to crash. But give the world a chance; it’s working on it.

Spurred by a new wave of Skype-linked families, Hulu-watching flash mobs, and HD-video downloaders, global internet traffic is likely to quadruple by 2012. That’s an internet 75 times larger than it was just five years ago. It will be generating 27 exabytes—nearly 7 billion DVDs worth—of data each month. Start stacking those DVDs on January 1, and you’d be at the moon by tax time.

Furthermore, most of those DVDs would be full of video. In 2008 alone, traffic generated by YouTube video was more than the whole sum of traffic crossing the U.S. internet backbone in 2000, according to Cisco. Right now, streaming video’s slice of the pie is estimated at 25 percent, and is expected to account for a full 90 percent of global traffic by 2012.

Creak though it may under the weight of new traffic, the internet is holding its own against this onslaught of new data. Internet service providers are simply ramping up their infrastructure upgrade plans in response to the traffic growth. As a result, international internet backbone utilization actually declined in between mid-2007 and mid-2008, according to the communications analysts at TeleGeography.  A big part of making this crush bearable is the continued development of content delivery networks by companies like Akamai. Their software and systems accelerate the delivery of their corporate customers’ Web content by putting copies of those files closer to the people requesting them. It's a little bit like delivering a truck full of Coke to your grocery store rather than insisting you fetch a cold one from Atlanta every time you get the itch for a soda, explains Akamai director David Belson.

Wireless broadband in the US is caught in a crossfire. Two new technologies, capable of delivering five to six times the speed of 3G networks currently used in the U.S. are battling for dominance. Verizon Wireless embraced one of the standards, called Long-Term Evolution, while Sprint Nextel choose the competitor, WiMax. Other mobile operators are simply fine-tuning their current networks to wring out every bit of speed possible without having to install the added equipment the newest technologies would require.  In the long run, improving broadband service in the face of a massive traffic increase comes down to a series of rather mundane infrastructure improvements. Last year, for instance, Comcast deployed new software that doubled download speeds for about 30 percent of its broadband customers. Internationally, homes in Japan, the Ukraine, and the Netherlands are  starting to get optical fiber connections capable of delivering gigabit-speed connections.

In the best-case scenario, the internet will continue to build out and keep working much as it does today: Not perfectly, but reliably great, and available to more and more people. Right now a new generation of undersea cables are being laid to, from, and within Asia, a massive series of projects that will both boost international connection speeds and reduce network latency—the time delay between when a packet of data is sent and when it arrives—worldwide. Similarly, low infrastructure areas will soon be able to have fiber-fast internet service delivered by satellite. By the end of 2010, O3b Networks expects to have as many as 16 satellites circling the equator, providing low-cost broadband to 3 billion people living in underdeveloped nations.

In the worst-case scenario, the crush of people and data turns the internet into a choppy sea of unpredictable spikes in demand, faltering whenever a virtual event draws a crowd or a malicious hacker decides to attack one of the protocols responsible for shuttling data around the globe.

Most likely, the future of broadband will be a little bit of both scenarios. According to Akamai’s latest survey, the No. 1 source of internet attack traffic, globally, is the USA, with China a close second.

Still, according to Shankar Sastry, dean of the College of Engineering at the University of California– Berkeley and director of UC Berkeley’s Team for Research in Ubiquitous Secure Technology (a.k.a. T.R.U.S.T.), the biggest challenge to the advancement and viability of an open internet is its most ordinary users. As long as human nature is what it is, the internet will be vulnerable. People lose laptops, pick easily guessable passwords, and open attachments from strangers. Fashioning solutions will mean changing the way the Internet works—and the way we work with it—little by little. Sastry calls the internet “a utility, no less hallowed than the traditional utilities,” and would never claim it could be as fast and invulnerable as it can be in the lab. But they’re working on it.