|
International internet traffic growth defies pessimism
 Honolulu - special to the IIC.Time was, telecom sector growth was a macroeconomic feature and heavily predicated on overall GDP growth, nationally and globally. Recessions in turn meant a fall in demand. In many cases, they still do so. But if that is the case, no one told the Internet that the world is still going through a major economic downturn. “We have seen no such recessionary effect", says Stephan Beckert, Vice President of Telegeography, a Washington DC based consulting firm specializing in traffic and bandwidth analysis. “[Traffic demand]has grown around 60% annually in each of the last three years,” says Mr Beckert. “So it doubles - at least - every eighteen months.”
The prodigious demand means that frequently service providers - such as the major telecom companies around the world - scrabble to deploy and buy new capacity. They are continuing to do: Mr Beckert calls the near-universal capacity expansion “phenomenal”.” The vast majority of the international backbone capacity that supplies the internet is carried on fibre optic submarine cables laid on the sea and ocean floors between countries and continents.”[Service providers] are now adding more capacity annually that was actually in service in total as recently as 2007”, he says.
But balancing supply and demand is a delicate and occasionally, hazardous, game for the backbone industry everywhere. In the last great upswing, the dotcom boom a decade ago, suppliers and carriers alike rushed to deploy systems for which there was ultimately no demand. The industry collapsed under the weight of the bad investment and took more than five years to regain its equilibrium.
It was a searing experience for those involved. Key routes such as the trans-Atlantic and, particularly, the London-New York segment saw such overbuild that prices offered to customers plummeted and Chapter 11 was the only recourse for several of the system operators involved. Not only that, but technology continually found new ways to expand the available capacity of systems actually in use so delaying any recovery through new system demand.
But, with Internet growth remaining at such levels, the game may have changed again. In fact, in each of the last two years, the industry has laid around 15 or 16 new cables, mostly relatively short haul systems designed to fill in pockets of high demand and cater particularly for emerging markets in the Middle East, South Asia, Central America and around Africa, which was the last great gap left after the previous boom. Meanwhile, Mr Beckert says that in many places, there is still a substantial amount of capacity left on many systems but this could disappear quickly. Even on the fated trans-Atlantic route this may apply: the aggregate potential capacity overall across the ocean is 53Tbps (Terabits per second) but currently only 12 Tbps is lit or active and available for traffic. Of this, 10Tbps has been purchased commercially and of this, only around 4.6Tbps is actually being used for real traffic. Telegeography estimates that some 77% of traffic is based on Internet demand with the remaining coming from private network customers.
“Considerable potential capacity exists on many routes for now”, says Mr Beckert but the industry is already asking, ‘When will we run out?’ If the eighteen month equation still holds, he says, and there were no increase in the upgradeable capacity as currently estimated, Telegeography says exhaustion would be somewhere in the 2014 timeframe for both Atlantic and Pacific routes. But he says that upgradeability is a ‘moving target’ given the rate of technological innovation in research laboratories. But, carriers are increasingly needing route and system diversity given the accidents that befall the systems on the sea bed. “There have been three major multiple cable faults since December 2006 and this has really drove home the need for diversity,” he concludes.
- Stephen McClelland
Only 595 days to go
 Honolulu - special to the IIC. “The countdown clock is now around 595 days,” says Martin Levy of Hurricane Electric, a US ISP, having just "jumped down" because of a wave of fresh IP address allocations. He adds ominously, “the clock has gotten to the point where, if we are standing here in two years time, the conversation will be very different.”
Yannick Pouffary puts it no less dramatically: “The perfect storm time has come”, she says.”I’ve been talking about it maybe for 10 years…but even though there is a major storm, people don’t learn.”
The incident they both fear is the consequences of the exhaustion of available address space on the Internet. To date, the relevant provision - IPv4 - has been able to accommodate everyone and everything but the explosion of usage means that it is predictably due to run out at some stage in the near term. Capacity is reportedly down to its last 10% and falling. At the zero point, no further addresses can be issued and as the OECD puts it, the danger is that this will occur at a global level with the ability to choke expansion and innovation, and be critical for the future of the Internet economy. It is likely to affect all businesses that depend on IP addresses for their growth, says the OECD.
Waiting in the wings, or at least in transition, is IPv6. It is designed to take to the heavy lifting of the next stage of Internet evolution. Making the transition, say its proponents, will not only deal with the address problem but provide a framework for other requirements as well. “IPv6 is designed to solve many of the problems of IPv4 such as address depletion, seamless mobility , security, network plug-and-play issues and extensibility," says Ms Pouffary, an acknowledged industry expert deeply involved with the industry association at the centre of the issue – the IPv6 Forum. And she is clear about the ultimate pressures:-"the convergence of everything over IP is the real driving force behind the move to IPv6."
The switch offers significant advantages in network architecture and the address space on offer is effectively unlimited. The problem is that moving globally the entire communications industry and its myriad players to voluntarily plan and adopt it is a mammoth task. It may well be the largest single initiative every attempted in the industry, with only Y2K being remotely comparable. Everyone from vendors to service providers to enterprise network operations at every level from network routers to handsets is potentially affected. But it is the transition that continues to cause the problem. There has been plenty of notification and over the past 10 years: NTT, Japan's incumbent phone company, has been working out IPv6 for 16 years and has fully implemented it. But whilst many players have already implemented IPv6 and moved on, there are significant numbers of otherwise sophisticated players that do not seem to have IPv6 “on their roadmaps” says an expert. The lack of insight, says the IPv6 Forum, seems to extend to people at the top of the industry. The result could potentially be a mish-mash where one day a significant number of users will simply be denied the Internet.
The prospect seems real enough to the IPv6 fraternity. “We are going to be in trouble unless we react to [the problem]," says Ms Pouffary, "we need to face [the prospect of] the Balkanization of the Internet.” And the accompanying issues are extremely serious as well, she says, in particular, security. “Unless we deal with security we will face a major disaster,” she states flatly and she says this will be a business continuity issue.
 But meanwhile, the issues involved say almost as much about what the Internet is and how it evolved. “The internet to date has been a grass roots effort,“ says Latif Ladid, President of the IPv6 Forum. “When you get to the next version, things get tough. You need a top level approach to make it happen.” He frankly admits:”the task is huge.” Governments are involved in various IPv6 task forces and programmes and clearly keen to see a stable coherence in the industry practice. Meanwhile, for everyone else, there is sustained evangelism: “Our next job is to make sure the vendors hear us,” says Yannick Pouffary.
- Stephen McClelland
Highway, letter post, or simply, utility?
 Honolulu - special to the IIC. If broadband is the infrastructure of the future, then what should a deployment model be? More precisely, how far should the public sector involve itself directly in the creation of networks to un-served or under-served communities. It’s a recurrent question and there are as many answers as there are players.
A wide spectrum of both opinions and options exists, especially after the financial crisis, which variously believes governments should simply mandate and pay for advanced broadband networks to maintain social cohesion and national competitiveness to the opposite extreme where any government intervention is seen as questionable and a market distortion. In this view, a market given a free hand and a competitive environment would inevitably deliver broadband at the greatest consumer benefit possible.
The question is, perhaps, most acute in the US where attitudes traditionally reflect the notion of free market capitalism. But as critics of past US policymaking point out, the US ranks only 15th on the most commonly accepted scale of international comparisons of broadband deployment produced by the OECD. They warn, in turn, this lagging performance will have wider implications for the future of the national economy. In fact, the US is also home to probably the most numerous community of municipal networks in the world; around 200 have been deployed over the last few years, but there is no single working model of what they do or how they should operate, says Chris Walker, Director of Operations at NoaNet, a non-profit service provider in Washington state that has deployed some 3400 miles of fibre optic network. Invariably these networks operate at a wholesale level in the telecom ecosystem although there is political pressure to allow them to supply retail services and for the first time compete with the major players in the industry, the telecom carriers.
Comparable operators exist throughout the United States but the approach is invariably a patchwork one, and the market complicated because of the presence of large incumbent carriers and cable TV companies. It’s a big country and the potential costs of a new broadband infrastructure are correspondingly large. Various estimates put the cost of a national 100Mbps fibre optic network at around USD 350 billion. It would be significantly cheaper to deploy one operating at only 50Mbps and cheaper still for lower speeds.
These are thought experiments because the idea of a nationwide publicly-funded system that would achieve this seems completely off the table in the US at least for now –although various officials in the Obama Administration are said to be highly impressed with the great sweep of Australian plans for a multibillion dollar broadband network and what may well be a structurally-separated incumbent operator. The FCC seems less than enthusiastic on the separation part however.
Meanwhile, the hunt is clearly to find the best policy model to stimulate broadband deployment by the private sector. Invariably the hunt is equally for a metaphor of what the desirable outcome would look like. Just how does broadband compare with other initiatives? Even in the US, there has been significant economic pump priming, notably in the early days of the interstate highway system, to kick start economic development and improve employment. Before that there was a national, flat rate postal system, and of course, utilities that expanded over time providing everything from power to water.
But a specially convened PTC session remained sceptical in general about both the benefits of government-sponsored broadband and even if it could be categorized on this basis, except in the most general of metaphors and many executives seem to suggest that government involvement is at best marginal and at worst, either wasteful or downright distorting in terms of marketplace behaviour.
But executives say that current US thinking suggests a lack of enthusiasm for the pump priming and pilot projects that have been publicly funded. Some, such as Todd Bundy, an executive at ADVA Optical Networking, accuse the US Government of promoting a ‘silo’ mentality when it comes to network architectures in even running such projects, such that one application is considered in isolation from any other possible uses. “I’d like my tax dollars to be spent in the most efficient way possible,” he adds.
- Stephen McClelland
|
