2013-11-26

In April 2009, Australia’s then prime minister, Kevin Rudd, dropped a bombshell on the press and the global technology community: His social democrat Labor administration was going to deliver broadband Internet to every single resident of Australia. It was an audacious goal, not least of all because Australia is one of the most sparsely populated countries on Earth.

The National Broadband Network (NBN), as the project is known, would extend high-speed optical fiber directly into the homes, schools, and workplaces of 93 percent of Australians. The remaining 7 percent, living out of fiber’s reach in rural areas and remote pockets of the vast outback in the middle of the continent, would be linked to the Internet via state-of-the-art wireless and satellite technology.

Governments and telecom carriers in other countries, such as Japan, New Zealand, Singapore, and South Korea, have similarly embarked on endeavors to deploy widespread fiber-to-the-premises (FTTP) networks. But those countries are much smaller and more densely populated than Australia. The country has roughly the landmass of the contiguous United States but only 7 percent as many people—fewer, in fact, than the state of Texas. To lay a nationwide fiber footprint, the government would need armies of workers and unprecedented access to rights-of-way, utility poles, and underground ducts.

And indeed, the NBN’s estimated cost was high: The latest figure was AU $45.6 billion (US $44.1 billion). It would be one of the largest, most pervasive FTTP rollouts any government has ever attempted. But although the price would be great, so would the impact: The network would bring broadband access to underserved areas, but it would also raise standards of living everywhere by driving innovations in telemedicine, remote education, e-commerce, and e-governance. A government-funded report released this year by Deloitte Access Economics concluded that the NBN would provide job opportunities, time savings, and other benefits worth, on average, AU $3800 (US $3600) per household per year by 2020, when construction would be nearly complete. In addition, fiber’s enormous bandwidth capacity means that transmitting and receiving equipment could be upgraded indefinitely at low cost, allowing the NBN to keep pace for decades with the incessant demand for higher data rates.

Yet despite these benefits, some conservative politicians and media outlets vehemently opposed the plan. In the campaign leading up to a national election this September, the fate of the NBN was vigorously debated. Although polls showed that the majority of voters supported the project, they nevertheless rejected the Labor Party and ushered into power an alliance of moderate conservative parties known as the Coalition, whose leader and now prime minister, Tony Abbott, promised to drastically scale back the national network.

So now, after three years of planning and construction, during which workers connected some 210 000 premises (out of an anticipated 13.2 million), Australia’s visionary and trailblazing initiative is at a crossroads. The new government plans to deploy fiber only to the premises of new housing developments. For the remaining homes and businesses—about 71 percent—it will bring fiber only as far as curbside cabinets, called nodes. Existing copper-wire pairs will cover the so-called last mile to individual buildings.

Such issues are not unique to Australia. Enthusiasm for near-universal broadband was once widespread, and it is still being pursued in the countries mentioned above, among others. But the ardor has cooled in recent years as legislators in many parts of the world move to cut government spending. Unfortunately, as is so often the case with technology, the public debate is beset by misunderstanding, misinformation, and a general lack of technical knowledge. A rare opportunity for growth and development is about to be lost, and disappointingly few people fully grasp the implications of that loss.

For example, in Australia, the Coalition is pursuing a fiber-to-the-node (FTTN) strategy because it would be much cheaper in the short term—about two-thirds the price of the original NBN. But that calculus overlooks the longer-term realities. Copper links simply lack the capacity to support the massive growth in data consumption that analysts predict. Eventually, Australians will have no choice but to replace those links with fiber, probably before the end of this decade. At that point, upgrading to an FTTP network will add to the cost of the FTTN rollout, increasing the total investment beyond the price of installing that fiber today.

And in delaying the deployment, Australians will have passed up a unique chance to become leaders in the global digital economy—an opportunity they may not get again.

Australia in the Slow Lane

Today in Australia, as in much of Asia, Europe, and North America, commercial carriers own and operate competing landline networks. Such an arrangement normally encourages carriers to stay at the forefront of technology. However, it can have disadvantages as well: In a thinly populated country such as Australia, carriers may cherry-pick customers in the few dense urban centers where they know they can make the most profit. Consequently, progress is slow to reach the vast majority of people living in rural and suburban areas.

It’s not surprising, then, that among developed countries, Australia is notable for its paucity of fiber-optic links. The highest rates are in Japan and South Korea, densely populated countries with small landmasses, where fiber accounts for more than 60 percent of broadband lines. In larger, more sparsely populated countries, such as the United States and Canada, rates are much more modest. In Australia, the rate is less than 2 percent.

Today, more than two in three Australian households have fixed broadband subscriptions. Most of those connections still use digital subscriber line (DSL) technology, which transmits data packets at higher frequencies than do analog voice signals, enabling Internet traffic to travel over telephone lines at relatively high bit rates. In a DSL system, twisted copper pairs, also known as loops, connect each customer to a central switching office. There, a rack of modems known as DSL access multiplexers, or DSLAMs, link the local loop to the backbone networks of various Internet service providers.

The problem with relying on DSL for broadband service is that many modern applications, including ultrahigh-definition videoconferencing and 3-D television, already require faster transmission speeds than these lines can provide. The biggest bottleneck is the copper itself. Due to the electrical properties of the metal, signals distort and weaken considerably with distance and can interfere with signals traveling through neighboring wires. This severely limits the bit rate of connections, particularly long ones. While customers close to a central office can receive rates as high as 24 megabits per second (using a common standard known as ADSL 2+), more distant customers experience much slower speeds. In Australia, where loops can be quite long and where some users opt for low-speed plans, the average Internet connection is just 4.8 Mb/s. And because the upload rate for DSL rarely exceeds more than one-fourth the download rate, the service doesn’t work well for high-bandwidth two-way applications such as videoconferencing.

The leaders of the Labor Party weren’t the first Australians to recognize the need for a faster, more inclusive network. Telecom carriers and federal advisory groups have been kicking around proposals for a national broadband network since about 2003. But it wasn’t until December 2007, after the Labor Party won majority power, that the government committed to the venture.

At first, Labor representatives thought the new network should use an FTTN architecture, which would require removing DSLAMs from central offices, located kilometers from customers, and installing new ones in nodes as close as a couple of hundred meters. The nodes would connect to the central offices via fiber and relay data to and from each customer’s premises using very-high-bit-rate DSL, or VDSL, the highest-speed DSL standard available at the time. These shorter copper loops would boost average speeds considerably—to as high as 50 Mb/s, depending on the distance between the node and the premises. The resulting FTTN network wouldn’t be nearly as fast as a full-blown FTTP grid, but the anticipated cost seemed more reasonable.

The government also assumed that the best way to build the network was to award the job to a commercial carrier through a bidding process. It would grant the winner a monopoly license and pitch in AU $4.7 billion to subsidize the cost of construction. Six carriers, including the market leader Telstra, submitted proposals by November 2008. To evaluate them, the government appointed an expert panel; I was among its seven members.

After studying the proposals, we agreed on two key points. First, we found that the global economic recession, sparked by the bursting of the U.S. housing bubble in 2006, was preventing Australia’s carriers from raising enough capital to fully fund the construction of a national network. In fact, none of the bidders came up with a viable business model. It was clear that unless the government bankrolled the majority of the cost, a commercial network would not likely succeed.

Our second observation was that an FTTN layout would be a bad idea. Using VDSL, a home connection could theoretically deliver 50 Mb/s, but only if the node sat very close to the house—a mere 100 meters or so away. Since the panel disbanded, a newer standard, VDSL2, has emerged. When combined with a novel interference-reduction technique called vectoring, it can provide download speeds up to about 100 Mb/s over short distances. And now an even faster standard known as G.fast is in the works, which promises download rates up to 1 Gb/s, but again, only for very short connections. For customers on longer loops, telecoms would be able to guarantee only about 50 Mb/s.

Market analysts project that data usage from a single family or small business could easily surpass that rate by 2020, and to meet this demand, Australia would need an FTTP network. Laying a cheaper FTTN footprint first would make little sense because it’s not a necessary step toward realizing an all-fiber system. In fact, an FTTN network requires special equipment and infrastructure, including nodes, that would have to be removed and discarded during an FTTP upgrade. An interim FTTN rollout would consequently end up costing Australians more in the long term than simply investing in FTTP technology today.

For these and other reasons, we recommended that the government itself create a national FTTP network. Incredibly, it accepted our advice.

Two Faces of the NBN

In April 2009, following Prime Minister Rudd’s landmark announcement, the Australian government established NBN Co to build and operate the future National Broadband Network. The governmentowned company would be responsible for connecting every home and business to more than 100 hubs, called points of interconnect, around the nation. These are places where commercial Internet providers and other content-delivery companies, called retail service providers, would hook into the network. To reduce some of the cost of laying fiber lines, NBN Co would pay commercial carriers to access existing underground ducts and pits and decommission copper telephone lines and DSLAM equipment. Telstra currently owns the vast majority of this infrastructure, and the government had agreed to pay AU $11 billion to access it.

As the sole owner of the new national network, NBN Co would run what’s known as a Layer 2 network. It would offer commercial providers a choice of speeds at set prices (from AU $24 per month for 12 Mb/s downloads and 1 Mb/s uploads to AU $150 per month for 1 Gb/s downloads and 400 Mb/s uploads). It would route the data to and from the providers’ customers using Ethernet protocol. The providers would add on the remaining layers, including data packaging, encryption, and error correction, and bill customers directly. Although NBN Co alone would manage the physical infrastructure, including the modems in people’s homes, providers could still compete, based on the type of content they offered and the quality of their service.

To construct the network’s fiber web, engineers opted to use passive optical network (PON) technology, a standard approach for FTTP networks. In NBN Co’s PON system, a single fiber would ferry data from a central office to a small curbside cabinet, where a beam splitter would divide the signal, guiding the light through up to 32 branching fibers, each leading to a separate premises. Unlike active optical networks, which electronically switch data at the cabinet in order to route it to its final destination, PON systems broadcast to all premises on a splitter. They rely on electronic switches at each customer’s terminal to weed out the neighbors’ traffic and encrypt the data to prevent eavesdropping. PON systems also tend to be cheaper, use less power, and are easier to maintain than active ones because they don’t require engineers to install and tend to switching equipment in outdoor cabinets.

When construction began on the NBN in 2010, the fastest equipment available for transmitting data on a PON network relied on an industry standard known as gigabit PON, or GPON, which can send 2.4 gigabits per second to each splitter. This overall capacity would be divided among all of the premises on a splitter. However, if several customers in a neighborhood opted for fast services, NBN Co would simply install more splitters at the cabinet—a quick, 20-minute job. This way, NBN Co could guarantee that every fiber-connected Australian who wanted the maximum 1 Gb/s rate could get it.

Inevitably, though, some people would fall outside this fiber footprint. About 7 percent of Australians live in rural communities or remote outposts where wired broadband access is technically or economically unviable. NBN Co would connect about half this population via fixed wireless towers equipped with standard 4G LTE technologies capable of delivering download speeds up to 25 Mb/s and upload speeds up to 5 Mb/s to each customer. The other half would be served by two new high-bandwidth geostationary satellites due to launch in 2015, which would provide similar data rates.

But no matter the type of access technology—fiber, wireless, or satellite—NBN Co would still charge commercial providers the same wholesale rates to use its pipes, ensuring equal and fair prices to all consumers regardless of location.

Many politicians and industry executives praised the NBN plan. Alan Noble, Google Australia’s head of engineering, called it “the greatest enabler of innovation.” Others said it was “a critical part in the evolution of the Internet” and “too good an opportunity to miss.” Nevertheless, the plan was controversial from the outset. Members of the conservative Coalition, concerned about rising costs and construction delays, have described the NBN as a “dangerous delusion,” a “white elephant on a massive scale,” and a “shockingly misconceived, wasteful exercise in public policy.”

Some of the early criticisms, particularly from media commentators, stemmed from technical misunderstandings. Opponents of the FTTP approach, for instance, often reasoned that the popularity of mobile gadgets is causing wireless technologies to advance so rapidly that they will eventually offer greater speeds than fiber, making the NBN obsolete.

The fallacy of this assumption is immediately apparent to anybody with a basic knowledge of wireless networks. Such connections will always be limited by the bandwidth capacity of a cellular base station, which must be shared among all its users. Even if one station could use all available radio spectrum to serve one customer, the bandwidth of frequencies that can be passed through an optical fiber would still be some 20 000 times as great.

What’s more, mobile systems may not be able to sustain their awesome growth without an extensive fiber network. Already, operators are deploying miniature base stations known as small cells in homes, businesses, and busy urban centers, to help expand capacity and bring services to places where traditional towers may not reach, such as indoors. The glut of data flowing through these cells will need to be hauled to and from an operator’s core network—a job that suits fiber very well.

Other critics of the Labor Party’s plan worried that giving NBN Co sole ownership of Australia’s physical network would stifle infrastructure competition, keeping prices high for consumers and slowing the adoption of new network technologies. This argument might be persuasive in more densely populated countries such as the United States, where high consumer demand usually ensures vigorous competition based largely on technology. Indeed, in the United States, Verizon began offering its FiOS FTTP service in 2005, and plans are now available to more than 18 million homes, 5 million of which have subscriptions, the company says.

But in Australia, providers have already demonstrated that a free market hasn’t produced good access options for most consumers. In the 1990s, for instance, Telstra and its competitor Optus strung separate hybrid fiber-coaxial lines, a faster service than DSL, to the same 2 million premises in some populous suburbs of Sydney and Melbourne. Meanwhile, millions more premises missed out on the upgrade.

By far the biggest concern about the FTTP model was, and still is, that the benefits won’t justify its high cost. The Coalition argues that an FTTN network, though less than ideal, would provide more value per dollar. But the numbers just don’t add up.

An FTTP network offering peak speeds of up to 1 Gb/s would have cost Australians about AU $3450 per premises, according to NBN Co’s cost analysis. By contrast, the new Coalition government estimates that each FTTN connection, capable of guaranteeing up to 50 Mb/s, will cost on average around AU $2320—a whopping two-thirds the cost of a vastly superior FTTP link. And if consumer data rates continue to climb as fast as analysts predict, many FTTN customers will probably want to upgrade to FTTP technology before 2020. To accommodate them, the Coalition government plans to offer “fiber-on-demand” service, in which a customer could choose to pay out of pocket for installing fiber from a curbside node to a home or business. These upgrades would likely add another AU $1000 to $5000 to the price of each connection, depending on the length of the fiber and the amount of labor required.

In the meantime, an FTTP network using GPON infrastructure could last well into the future. Upgrading it to the next-generation standard, called XGPON, which will support up to 10 Gb/s, would simply require replacing some of the equipment in central offices and the terminal modem at each customer’s premises—for a likely total bill of no more than AU $300 per connection. In the future, newer standards could provide even faster bit rates for a comparable cost.

It has been painful watching the formation of this “futureproof” network come to an end. I can’t help but think of the United States’ Interstate Highway System, championed by President Dwight D. Eisenhower in the 1950s, which paved the way, literally, for a booming transportation-based economy. In Australia, a fiber-based broadband highway could transform the country’s digital economy in much the same way.

Sadly, the new Coalition government seems impervious to these arguments and is determined to downscale the NBN. I am left clinging to the hope that Australians will realize the foolishness of abandoning the FTTP network and insist that their leaders reconsider or devise a new plan that’s not too far removed from the Labor Party’s revolutionary vision.

This article originally appeared in print as “Australia’s (Less Super) Super-highway.”

About the Author

Rodney S. Tucker is a professor of electrical engineering at the University of Melbourne and an IEEE Fellow. In 2008, the Australian government appointed him to an expert panel that recommended building a nationwide all-fiber access network, now known as the National Broadband Network. Although this ambitious project is now in danger of being downscaled, Tucker considers his work on the panel “one of the most significant contributions of my engineering career.”

Show more