When the FCC issued its National Broadband Plan earlier this year, it set some modest goals for the nation: 100Mbps to 100 million homes by 2020, universal service of 4Mbps everywhere.
Australia plans to do things… a bit differently. Within the next eight years, the Australian government will spend AUS$43 billion (US$38 billion) to build its own “world-class broadband infrastructure” that will deploy fiber to 93 percent of all Australian homes and bring 12Mbps broadband to everyone else. The network will be wholesale only and will be open access, enabling every ISP to use the fiber to offer services.
What about the presence of a “highly profitable vertically-integrated and horizontally-integrated incumbent with a monopoly position in most fixed-line access in many backhaul routes”? (This is a reference to Telstra, which controls most of the copper telephone line in the country.) Well, Telstra can partner with the government and make a bit of money off the rollout, or it can choose not to cooperate and the government will do what it wants anyway.
The message is clear: Australia refuses to be held back by the business decisions of a single company. As a new report on the Australian plan makes clear, there’s not much attempt to make this easier on Telstra.
“Existing participants will need to adapt to succeed,” says the government report.
The new report (PDF), out today from consultants at McKinsey and KPMG, is an “implementation study” to make sure the government’s plan is feasible—and the report is remarkably positive.
The government initially planned to spend AUS$43 billion (about US$38.9 billion) and build a fiber network that would reach 90 percent of Australian homes. The rest would get wireless broadband, while the truly rural would have access to a new class of satellite service that can deliver a least 12Mbps to the entire country.
Too modest! says the report. Instead, Australia should bump its fiber build up to 93 percent, provide fixed wireless service for four percent, and Ka-band satellite service for the remaining three percent—and it can all be done in eight years and on budget. (Australia will apparently launch two of its own Ka-band satellites to provide a redundant national connection serving up to 350,000 homes.)
“The new fiber infrastructure covering 90+ percent of premises will be long-lived and will redefine the current industry structure, superseding the copper network that has been in place for over 50 years in many places,” says the report “the fiber laid over the next decade will be in place for possibly as long again.”
As for Telstra, there’s not much consolation here. The new fiber buildout will produce a “fundamentally different industry structure… This change will accelerate the evolution of the industry. At times this may be smooth; at other times it will be uneven. New business models and companies will emerge.”
The incumbent can make money by doing a comprehensive deal with the government to “share infrastructure such as ducts, pits, exchanges, and unused backhaul fibers.” Telstra has over 100,000km of conduit in the ground already, and it’s estimated that 50-80 percent of these ducts still have unused space to run fiber. This could save the government significant money in trenching costs if Telstra will sell access. Likewise, the incumbent has plenty of dark fiber installed around the country to serve as backhaul links between exchanges; the government could also save the expense of laying its own fiber over these routes if it partners with Telstra.
But it’s clear that the new network won’t be held hostage to Telstra’s demands. The consultants conclude that, in the absence of an agreement, [the fiber network] should proceed to build both its access network and its backhaul unilaterally.”
“Fiber”: Australian for “broadband”
Because of its harsh interior geography, nearly all Australians live on the perimeter of the country. The report notes that only 9 percent of the country’s land mass is inhabited, and 90 percent of the people live in just 0.2 percent of the space. It’s relatively easy and inexpensive to run fiber to these locations, but it’s still a big job that will take years to complete.
Just think about what’s involved: fiber will run alongside almost every road in the country. The report notes that 250,000km of fiber optic cabling will be needed and that the new network will have to make “5,000 customer visits per workday” for eight years to hook everyone up.
The obvious downside of this approach is that no other wireline competitors will likely emerge to challenge the government-controlled fiber network. On the other hand, there’s not that much competition right now; only 20 percent of Australian homes can get cable service, and most others are left with lines from Telstra (which does resell to numerous ISPs, but it retains control of the access network).
To encourage competition, the government will not enter the ISP business itself. Instead, the plan is to wholesale fiber connections out to ISPs, and the network will be open to every ISP that wishes to participate. This is similar to what Google plans to do with its 1Gbps testbed fiber network somewhere in the US—but over an entire country.
The report expects that wholesale prices should start at AUS$30-35 (US$27-32) per month for 20Mbps broadband with voice service.
Australia’s plan may never have worked for the US: we have much stronger duopoly competition than Australia, and a company like Verizon has already invested $18 billion-20 billion in fiber. Also, population density is different between the US and Australia. Still, it’s an exciting project to watch, and it certainly makes us wonder what the US would look like with fiber to 90+ percent of homes and a healthy ISP competition over that fiber. (Google’s project seems designed to make US residents think about the possibility.)
But for now, we watch. We wait. And we hope our own rural residents are happy enough with their 2Mbps connections—when they can get broadband at all.