Is the internet doomed?
All this data has to pass over a network that was designed for phone calls and faxes, not IPTV and video-conferencing. Bottlenecks are likely as investment in network capacity struggles to keep pace with new services. According to Brett Swanson of the Discover Institute, this is most likely to affect the “edges” of the internet – the final mile or so of cabling between telephone exchanges and homes – but predicts that the rest of the infrastructure is “totally unprepared” for the coming “exaflood”.
The tsunami-sounding “exaflood” refers to an exabyte, or one thousand million gigabytes of data. Until recently, an exabyte had simply been a mathematical concept, but Cisco estimates that total internet traffic will reach 10 exabytes a month this year. To make the infrastructure fit for purpose, the researchers at Nemertes believe the US alone would need to spend $55 billion on upgrading the networks, while it would cost anywhere between £7 and £20 billion to lay next-generation fibre-to-the-home (FTTH) in the UK.
It all sounds ominous, but is crisis really imminent? Much of this infrastructure conundrum centres around one of the most profound debates of the internet’s short history, the dispute over “net neutrality”, which could shape the future of the web and dictate the way we pay for data.
At the moment, all data packets are treated equally; most ISPs – assuming you’re within your usage allocation – simply feed the data requested by your clicks with little interference. The big carriers act in the same way, so the companies running the transcontinental cables just pipe the video download you requested from an Australian server, for example. Carriers have traditionally built excess capacity into the system to ensure that, even at peak times, there’s plenty of bandwidth to go around.
The carriers, however, want to change this state of affairs and take control over how much data crosses their networks. They’d prefer a tiered internet, where telcos could charge either content providers or consumers for priority access, so that streaming video comes through in its best possible quality, for example. Other content – perhaps from rival broadcasters or ISPs – would be deprioritised.
Why, argue the two-tiered internet advocates, should telcos invest billions in building new infrastructure when they can only make the most marginal return on investment, while parasite organisations such as the BBC or Skype reap all the rewards? Another industry report, from Rensselaer Polytechnic Institute, says a neutral network means telcos have to spend 60 to 100% more on bandwidth to ensure spare capacity, which they wouldn’t need to do in a tiered environment.
The war around “net neutrality” is being waged in the US, although its aftershocks will certainly be felt here, and this fundamental dilemma forms a backdrop to many of the reports and arguments currently predicting the demise of the web.
Yet, such reports deserve to be treated with a healthy dollop of scepticism. One of the less-reported aspects of the Nemertes “internet brownouts” study was that it was funded by the Internet Innovation Alliance, an industry and non-profit lobby group aimed at changing the way the net is regulated in the US. Among its members are AT&T, Nortel, Level3, Alactel-Lucent and Corning, all of which have major interests in drawing attention to the fact that traffic is rapidly rising across their networks.
“ISPs are terrified with the unrestricted growth in P2P traffic,” says Aydin Kurt-Elli, CEO of business broadband supplier Lumison. “Hence the debate about allowing carriers to start doing paid-for prioritisation of traffic. Indeed, the conspiracy theorists might argue some of this discussion regarding capacity running out is driven by the same agenda.