Andrew Ferguson July 12, 2013
Labor's fast broadband plan is superior to the Coalition's because it looks to the future.
Buried optical cables, unaffected by radio interference and weather damage, provide the best reliability for broadband access. Photo: Nic Walker
Offered as a path to the future, and derided by some as a costly extravagance, the plans for the national broadband network may be re-evaluated under the new ALP leadership. In a recent opinion article in The Sydney Morning Herald, Frank Zumbo of the University of New South Wales questioned the benefits of supplying fibre to the premises (FTTP) instead of a (potentially) cheaper solution such as ubiquitous wireless networks, or a simpler fibre to the node design, such as that proposed by the Coalition.
Both of these alternatives stop short of connecting individual households with future-proof fibre optic connections.
I find Zumbo's comments about the national broadband network troubling. I will leave the political and policy questions of cost, competition and construction to others. Instead, as a computer scientist specialising in computer networks, I will focus on the technical reasons why fibre to the premises is superior to other approaches, and the best solution for Australia.
Recent criticisms of FTTP include that it provides bandwidth which only a small portion of ''prolific'' downloaders will need and that it does nothing to improve mobility and convenience for wireless computing.
The criticism of bandwidth misses the transformative aspect of nationally deployed FTTP. Throughout the internet's history, increasing the speed and availability of broadband has consistently led to new, previously undeveloped, applications. At its start, the internet was a text-only system using slow telephone links. Over time, it was augmented with graphics on the world wide web, new applications such as Skype and other Voice-over-IP (VOIP) services, and in the recent past, video-streaming sites such as YouTube and Australia's Quickflix, as broadband connections became more ubiquitous.
By deploying more bandwidth in more places than needed today, the NBN will open the door to further advances in the home such as telemedicine, remote car and appliance diagnostics, and detailed security and environmental monitoring, applications popularly known as ''The Internet of Things''.
On the issue of mobility and convenience, the deployment of ubiquitous FTTP has the potential to improve wireless mobility, convenience and coverage just as much as other solutions. A fixed-line NBN will allow mobile phone operators to pay for the use of ''femtocells'' in subscribers' homes. Femtocells are essentially miniaturised cell phone towers, and can provide greater availability of high-bandwidth cell phone signals - at lower cost - than traditional towers. Femtocells are already successfully deployed in the US, Europe, Japan, China, New Zealand and elsewhere.
In addition, networking researchers have recently developed new methods to safely, securely and fairly share bandwidth between users and visitors on home wireless networks backed by the same provider. These methods eliminate concerns that a fixed-line NBN could only be accessed from one's own home.
It's worth adding that the current FTTP plan is superior to the alternatives for additional reasons: large wireless networks are generally difficult to deploy successfully due to their inherent interaction with the physical environment. Interference from other radio transmitters, building materials, and home appliances negatively affects a wireless network's performance. Furthermore, these problems' intermittent nature requires expensive over-engineering to overcome and difficult processes to identify their causes. Lastly, transitions in wireless standards can necessitate the complete and costly re-engineering of existing networks.
Also, most wireless networks are shared among multiple parties. Not only does this lead to unfair service degradation when one party uses it more heavily or with substandard equipment, it also leads to shared service failures. One cannot pop over to the neighbour's to use a working connection when all homes nearby share the same broken link!
For these reasons, buried optical cables, unaffected by radio interference and weather damage, provide the best reliability for broadband access. Greater reliability would make health and safety applications more practical; one particularly exciting possibility is the use of fixed-line NBN connections as a backup for emergency communications during disasters that damage police and fire radio networks.
Finally, a fixed-line FTTP design naturally provides a better user experience than other approaches, even those claiming to provide the same bandwidth speeds. To understand this, we need a brief technical detour.
Our online experience is affected by the connection's speed and its latency. Latency is the time needed to transmit data from its source to the destination. Low-latency connections provide a good experience when using audio and video chat, online games, or desktop-like applications such as Google Docs or Microsoft Office 365. Because a network's latency is a function of both its physics and the system's load, an FTTP design offers users the lowest (best) latency. Copper connections are next best, followed by fixed wireless connections, and then satellite.
Fibre to the premises is the right approach for modern broadband networks and is being used by telecoms around the world.
The NBN's current design is a sound plan for renewing and transforming Australia's telecom infrastructure. Let's hope last week's leadership shakeup doesn't derail it.
Andrew Ferguson is a computer scientist. He works on data centre networks at Google, and is completing his PhD in computer networks at Brown University in the US.
