High Speed Down Under

An Australian national broadband network began as a 2007 election campaign promise. Liberal party Senator Helen Coonan, then communications minister, said it couldn’t be done. The Labor party’s telecom expert, Senator Stephen Conroy, thought otherwise. Now that Conroy’s party is in the majority, he’s taken over as communications minister, and Australia plans to invest AUS $4.7 billion (US $3.1 billion) in a national network. After numerous delays and worry that the government would not follow through, the project seems to be on track—telecommunication companies vying for the job are to submit their proposals by 26 November. The government plans to pick a winner in early 2009.

The Australian plan calls for a fiber-optic network. But because of costs and practicality, analysts say, it will likely include a mix of fiber, WiMax wireless systems, and satellite services. How much of each nonfiber technology the government will go for is the big question that bidders are struggling with.

Building a national broadband network is a daunting task for any country, but particularly for Australia. The combination of a large land mass (slightly smaller than the contiguous United States) and low population density—just 2.7 people per square kilometer—make broadband infrastructure both a technological and economic challenge. Serving Australia’s scattered coastal cities—where 85 percent of the population lives—is one thing, but bringing broadband to the outback, where people often live kilometers from one another and hundreds of kilometers from the nearest city, is quite another.

The national network must deliver broadband to 98 ­percent of the population at a download speed of 12 megabits per second. Those targets might be hard to meet, given Australia’s current average speed of 1.7 Mb/s and a penetration rate of roughly 60 percent of households, according to a report by the Information Technology and Innovation Foundation, in Washington, D.C. While 12 Mb/s would be a huge improvement, it would still leave the country lagging behind others in its region, notably Japan and South Korea, where speeds of 63.6 Mb/s and 49.5 Mb/s, respectively, are average.

Two main bidders have emerged. One is Terria, a consortium of telecom companies including SingTel Optus, Australia’s second-­largest carrier. Terria will compete against Telstra, Australia’s principal telecommunications company. Telstra, in which the Australian government held a controlling interest until 2006, already has much of the infrastructure for a broadband network in place. Paul Budde, managing director of BuddeComm, in Bucketty, Australia, estimates that the Telstra fiber network could be expanded to serve all of metropolitan Australia at the required speed in early 2009. The challenge, he says, will be providing service to the suburbs, small towns, and rural areas.

Telstra has a clear edge, but in the event Terria snags the bid, it will push for an open network so that it can access Telstra’s infrastructure. ”From a national economic point of view—just like electricity, water, and gas—it does not make sense for that infrastructure to be duplicated or tripled to serve the same function,” says Ravi Bhatia, CEO of Primus, a company in the Terria consortium.

The bidders could not disclose the exact details of their proposals because of a government-issued gag order, but both bids most likely include a combination of fiber optics and wireless services such as WiMax, which, ideally, has download data rates of 10 Mb/s to 70 Mb/s, but in practice is more like 1 to 2 Mb/s.

The fiber part of the network, which is likely to be used only in the more densely populated areas, can come in two flavors, fiber to the node or fiber to the home. In fiber to the node, fiber-optic cables form most of the network. Then copper wires branch off from node stations—usually serving a few blocks’ worth of buildings—to service the individual customer. The use of a neighborhood’s existing copper greatly reduces the cost. But broadband speed over copper cables is limited by distance and the number of subscribers.

Fiber to the home eliminates the need for node stations but requires running new fiber lines to each building, something that makes economic sense mostly in greenfield developments, says Vince Pizzica, special advisor on technology strategy for Telstra. An entire network of fiber to the home would be extremely expensive. Mark Lutkowitz, a cofounder of Telecom Pragmatics, in Nashville, estimates that it could cost Australia as much as AUS $30 billion—well over the government’s AUS $4.7 billion subsidy.

Instead, analysts say, a combination of technologies makes the most sense. Budde says the best mix would be fiber to the home in all major metropolitan areas and smaller cities, with WiMax linking the homes up to 50 km outside the cities. For the 3 percent of the population living in remote rural areas, satellite is the only option, says Budde. Satellite would be a great deal slower than fiber and wireless—with download speeds between 800 kilobits per second and 1.5 Mb/s, and upload speeds of around 500 kb/s, says analyst David Gross of Telecom Pragmatics.

The Australian plan demands a fiber network, so it is unclear whether the government would support proposals that include a large wireless component. Peter Moon, managing director of Horizon Broadband Communications in Ballarat, Australia, says the government’s position seems to be shifting as it becomes evident that wireless is a more cost-effective solution. The government is thought to be expecting a proposal with 8 to 10 percent wireless, says Moon. But the general consensus in the industry is that wireless will compose 20 to 30 percent of the system. The government ”should have been technology-agnostic from the beginning,” Moon says. ”That would have been the smart thing.” Moon is hoping to have his antenna technology incorporated in the wireless component of whichever bid wins.

Once the bid is awarded, the winner will have to rush to roll out the network. Optimists still cling to an April 2009 time frame for most of the network to be up and running. But some analysts estimate that it will take as long as 15 years to deploy a full network.