The world's foremost internet traffic study and growth forecast, which historically has been proven very accurate, describes a further explosion of internet traffic around the world and in Australia. The findings illustrate a requirement for fibre optic cable "deep deep into the infrastructure" both for wired and wireless broadband connections.
The global study carries with it political ramifications in Australia where the opposition Coalition parties maintain that the optic fibre-based infrastructure currently being implemented by NBN Co is not needed to fulfil the promised benefits of the NBN and that suggesting so is "one big lie." However, the study further illustrates, using measured figures and reasonable growth curves, how the Coalition's alternatives won't just be unable to support the benefits to health, education, power distribution, business and society that NBN Co's current planned infrastructure will provide, but also that they won't be able to support the regular organic growth of the general internet requirements that we have now - within just four years!
Cisco's study
There are few (if any) more respected and influential broadband forecasts than Cisco's Visual Networking Index (VNI). Cisco is the company most associated with the actual infrastructure of the internet and is in the best position to measure the traffic that moves across it and, more importantly, the growth trends. These hold enormous implications for broadband infrastructure at international, national, regional, city-level and domestic scales and subsequently telcos and governments all over the world pay a great deal of attention to the study.
Different categories of Cisco's VNI report are released three times every year. They can be seen here for every country. (Note: this article ties together information delivered in two different briefings which is why some stats go to 2015 and some go up to 2016.)
For the past six years the forecasts have very closely resembled the measured figures - but ultimately they were conservative in their predictions. As such, the numbers coming from the last few forecasts, with their continued astronomical growth have huge implications for Australia's broadband infrastructure requirements with observations that broadband connection requirements will outgrow the current infrastructure by 2016.
Speaking at a recent VNI announcement was Dr Robert Pepper, Cisco's Vice President of Global Technology Policy. He has sat on the board of the Federal Communications Commission (FCC) in the USA and currently sits on the UK's equivalent, Ofcom. In these roles he briefs governments and network operators from around the world on infrastructure, what to expect from future data requirements and modes of broadband usage based upon traffic stats and growth curves. He is an American based in the USA and has no dealings with Australian politics. Some of the key points he made were:-
- That all roads point to the requirement of optic fibre being implemented deep into both wired and wireless networks.
- The future is indeed wireless, but it's mostly WiFi and not 4G.
- Wireless technologies need to be primarily methods of connecting to nearby fibre networks.
- That Australian mobile networks will soon have to join the US and UK in offloading data onto local WiFi networks in order to avoid congestion.
- That a 4G mobile user uses 28x more data than a 3G user.
- That new wireless spectrum needs to be opened up as quickly as possible to cope with growth.
- That as much wireless traffic as possible needs to be seamlessly offloaded onto the wired networks to avoid congestion.
- There is a huge increasing requirement for low-latency data transfer and high upload speeds.
- That a fibre to the node infrastructure which relies on a 'last mile' premises connection using Australia's current copper infrastructure, current HFC networks or fixed 4G-like wireless won't have the symmetry, contention ratio, bandwidth or latency to keep up with demand by 2016.
- That fibre needs to be very nearby every internet connection whether wired or wireless.
The report also goes into depth examining why some countries with good infrastructure have not yet made best use of it.
Glossary
Inevitably with an article like this there are some terms which can befuddle those not used to dealing with the technology industry. These are some of the most important terms to understand:-
WiFi - connecting to a nearby wired network, wirelessly. This wireless technology differs from the following...
Mobile network/3G/4G/macro network - are all terms used to describe wireless data flowing over the mobile phone network.
WiFi allows for super-fast connections to the wired internet from mobile devices like laptops, tablets, games consoles, set top boxes and phones but only at short range. The mobile network's range is far superior but data is far more expensive and performance drops dramatically when many people share the connection. It's typically used by phones, some tablets and some laptops at present.
Symmetry - the 'A' in 'ADSL broadband' stands for Asymmetric. This refers to the fact that virtually all standard consumer broadband connections have high download speeds but much slower upload speeds. For instance, a top ADSL2+ Australian connection might provide 12Mb/s download speeds but only 1Mb/s upload speed. Future requirements require more symmetry - fast download AND fast upload speeds.
Latency - Basically this is lag. If you've ever seen a satellite interview on TV you'll have noticed the delay between someone finishing speaking and the other person responding. That's because satellite offers a high latency. A low latency reduces the delay.
Global growth
Global internet traffic will increase 4x from 2011 to 2016 and break the Zettabyte barrier (1.3 Zettabytes) for the first time - that's 10x more than all internet traffic generated in 2008 (121 Exabytes) and is equivalent to one trillion gigabytes or 38 million DVDs per hour.
Global internet traffic has been increasing at 32% Compounded Annual Growth Rate (CAGR) year on year. The numbers are now so large that the "Law of Large Numbers" means that percentage rate increases might be going down but actual increases are still "dramatic."
In that time, the average broadband speed will increase almost 4x from 9Mb/s to 34Mb/s. When that happens, over two YEARS of video will be crossing the internet every second.
The dominant reason will be the growth in internet video.
Video
Short form (sub-seven minute) YouTube style clips made up 20% of the video figure in 2010 but, despite large growth, they are set to make up 14% of the figure in 2015 due to larger increases in other areas. The total amount of short form video consumption in 2015 is over half the total broadband consumption for 2011!
The fastest-growing category which, as Pepper says, was "almost non-existent in 2010" is internet video to TV. This is fuelled by production houses developing long form (over seven minutes) web only content. An example given by Pepper describes how, in the USA, the sports associations dealing with football, baseball and basketball have taken back control of their distribution rights: they're still using free to air broadcasts, but they are offering simulcast services online along with other packages to create multiple revenue streams through their own distribution.
In a similar vein, US video on demand service, Netflix, accounts for a staggering one-third of all US downstream broadband traffic. Last week, Paul Colley, Sony Group Manager for Network Services and Technology, described how all of Sony's new TVs now focus on displaying video from the internet because, "Most television will be consumed over the internet in three years." Sony's rivals have said similar.
It doesn't end there. Kevin Bloch, Cisco Australia's Chief Technology Officer, pointed out that new technologies consistently appear and skew the predictions further. "Tablets didn't really exist two years ago but now they're driving the market". Apple's move to HD displays in its latest iPad and latest laptops could impact figures further such is their enormous popularity and influence on content.
Locally, Australia's consumer internet video traffic grew 55% in 2011. In 2015, internet video traffic will be 81% of all Australian consumer internet traffic - up from 50% in 2010.
Good news for anti-piracy advocates is that File Sharing traffic is increasing the least. Pepper points out "What we're seeing globally is that when the video is made available in forms and packages that you can easily use and pay a modest price for, people will pay... 60% in 2010 [globally] will only be 24% in 2015... It looks like Hollywood has learned from the mistakes of the music industry."
The phenomenal success of Netflix in the US is testament to this - the service costs just $8 per month and gives unmetered access to vast catalogues of films, TV series and documentaries. However, because there are long (in some cases 10 years) and lucrative licensing agreements with 'overseas' traditional-media partners, the switchover to legal services that deliver up-to-date content, lags globally. Australia's "illegal" devouring of series like Game of Thrones is illustrative. It will be interesting to see how far Australia lags behind the USA in this regard come 2016.
Australia
In Australia, internet traffic reached 120 Petabytes per month in 2011 which was up from 93 Petabytes per month in 2010. This was in line with global trends. However, Australian internet traffic will grow 6-fold from 2011-2016 while the global trend is 4-fold.
By 2016 there will be 142 million networked devices in Australia. That's 5.7 per person - up from 4 in 2011. 40% of them will be wireless connected.
Fixed broadband speeds will increase from a 7.9Mb/s average in 2011 to 36Mb/s in 2016 (the US will hit 37Mb/s). Internet video traffic will make up 80% of Australia's consumer traffic by 2016.
There will be 20 million internet users in 2015 - up from 14 million in 2010. Most will not be connected by wire - WiFi, 3G or 4G. Dominant will be fixed WiFi.
A key figure was that in Australia the average internet user will generate 19.5GB of internet traffic per month in 2015 - up 562% from 3GB per month in 2010 (which compares to rises from 13GB to 47GB in the USA and Western Europe). More importantly, the average Australian internet household will generate 44.2GB traffic per month in 2015 (compared to 109GB in the USA and 99GB in Western Europe) up 588% from 6.4GB per month in 2010.
Pepper says, "Australia is about three years behind America and Western Europe in its consumption pattern." This is not entirely surprising considering the non-ubiquity of broadband access in Australia and that our market has only very recently started to remove strict data allowances on internet plans that didn't exist elsewhere in the world.
"It won't work... look at the figures."
Australian infrastructure requirements
These massive rises in local data consumption begged the question as to whether fibre to the node as favoured by the Coalition would actually work. Pepper's response to the question of connecting each home to the nearest node using 4G-style LTE fixed wireless, as has been mooted by the Coalition was telling. Without having any knowledge of the politically-charged question, and staring as though I hadn't been paying attention to anything in his presentation he said, "It won't work... look at the figures." He added that copper can be used for fixed connections in some circumstances but that it would have to be short, high-quality and with a low contention ratio - of around 8 connections per node. This rules out Australia's existing copper infrastructure and its HFC networks, which have much higher contention ratios (as well as slow upload speeds), for being able to cope with the growth in demand either.
A good analogy used here is that this is the equivalent of lobbying for a one-lane Sydney Harbour Bridge as that was all that was needed in the 1930s. However, while it took several decades to congest the Harbour Bridge and subsequent Harbour Tunnel the current copper network will likely be congested in just four years and fibre is the only technology capable of coping with the demand and future growth.
Cloud - (geek warning, feel free to skip over this bit)
Cloud computing has been a buzz word in the industry for some time. For businesses, storing information and running computing services from devices based on the internet is much cheaper than buying, maintaining, supporting and upgrading their own infrastructure.
Globally, cloud traffic will increase 12-fold between 2010 and 2015.
Pepper says, "Network requirements for each level of cloud application are usually 2-3Mb/s per application. (HD Video applications like telepresence are an exception and require 4-5Mb/s and a symmetrical connection.) So each household or business needs multimegabit symmetrical connections [to allow for] these applications to stack up and be used simultaneously. But advanced cloud services require less than 50ms of delay. No networks can currently meet this. The Japanese and Korean networks almost can.
"The FCC looked at differences between fibre and copper and found that fibre had about half the latency of copper (even short copper). If Cloud is growing this rapidly, and we have cloud as a core component of the use of internet broadband going forward, advanced cloud services [those requiring under 50ms latency] are going to require low latency which translates as fibre to the premise or fibre very very very deep in the network."
Pepper went on to point out that Latency was originally thought just to affect high definition video conferencing. But it's also about collaborating over networks and performing basic functions for small businesses and enterprise-level businesses alike. At a consumer level it's required for VoIP calls, Facetime, Skype, telepresence health and education services. These also require fast upload speeds.
Should anyone want some further reading on cloud, Parallels recently released its SMB (Small/Medium Business) Cloud Insights Report which describes how "the market opportunity for cloud services will grow to a total of $865 million in 2012." It goes on to say that "15% of Australian SMBs report using hosted servers, a 28% increase from the previous year. However, this is still below the global average of 20% of SMBs in other developed countries."
If you want even more information on cloud, there are many cloud-focussed business blogs on ABC Technology.
Mobile
Pepper points out, "The biggest trend that we see is called offloading. Between 65% and 85% of mobile phone data takes place indoors sitting down. Facetime won't work over the macrocell (mobile) network because it puts on such a load... that it won't work. It's partly to do with available bandwidth and part latency."
Pepper described some steps being taken to get round the problem, "AT&T built US public hotspots and signed a contract with Starbucks. The biggest trend that we see is called service provider offloading. This leaves voice on the macro [mobile] network and puts data on the fixed [WiFi] network... All towers are going to have to have fibre connections in order to accommodate multiple users and multiple connections."
Offloading is also happening in the UK with BT Openzone. There, phone and internet subscribers can seamlessly access BT's mesh of WiFi hotspots when out and about.
Pepper went on to say, "Offloading at home is important too" and spoke about Femtocells. These are being given away in some countries. They're already appearing in Australia. They sit in a home or office and your phone automatically connects to them like a regular mobile network. They boost mobile signals for those with weak reception and automatically channel data traffic off the network and onto your wired broadband connection. With the massive increases in data being consumed by 4G users, these may soon become more common in Australia.
Why is offloading so important? Apparently the following graph shocks every telco that sees it:-
Pepper says that the 2015/16 jump represents 4G adoption. The jump in data between 2015 to 2016 is 3x the total amount of global mobile data in 2012. He points out that what's driving demand is an increasing number of users, an increasing number of devices, massive video growth and increased speed demands.
"In the first year of usage, the average 4G user consumed 28x more data than other users. An artefact of early adopters? Maybe. But history shows early adopters have always led where others followed."
Back in Australia...
[Note: the sub heading should read Australian mobile traffic will increase 14x from 2011 to 2016 - typo]
Pepper points out, "Australia currently consumes 3x the mobile data of in China. But the China market will grow more quickly due to an increasing number of connected people and devices."
The breakdown for consumer versus business is as follows:-
What devices are we using? :-
How much data are we consuming on them?
By 2016 we'll be consuming 2.5GB per month on smartphones. 4GB on tablets and 7GB on laptops. Apparently the new (high definition) iPad may skew things further. Pepper says, "Apple keeps blowing up our models every year and [due to the 3rd gen iPad's penchant for HD content] it could be higher."
He adds, "In 2011 90% of mobile devices were laptops. Now 3G dongles (broadband connection modems) are being replaced by smartphones. Smartphone growth is 200%."
What will we be doing with our devices? :-
A sobering stat for Australian telcos is as follows: In Australia, by 2016, 59% of mobile users (11.8 million people) will be generating more than one gigabyte of mobile data traffic per month. In 2011, 0.9% of mobile users cracked a gigabyte.
Pepper says, "That's a huge increase. We show this to the operators and they kind of go [gasp!] How am I going to do that?" He continued, "Now imagine you're a telco in 2012 wondering what sort of investment is required to allow your network to cope with the increase in demand. New technology and business plans are required. Two thirds of the traffic is video."
He adds, "Furthermore, two-thirds of the mobile data we use will be sitting in the cloud":-
More alarmingly, after examining the first year of 4G traffic in the USA, the average connection generated 2.8GB per month which is 28x more than non 4G traffic which averaged 86MB per month.
Of particular note is the following. Pepper says he talks to governments all over the world. "Is it going to be fibre or 4G? The answer is both. You're going to need fibre to connect all those base stations... a lot of industry people aren't waking up to it."
He goes on, "Telcos used to think WiFi was their enemy. Now it's their friend." Kevin Bloch continued, "We spoke to Telstra who said that they didn't need to think about this as their 3G network was the best in the world. But it's all changed in the last 12 months."
Pepper gives an another example of our reliance on wireless, "Printing from an iPad in an office could send a 50MB document all the way back to the exchange and then back to the printer [which is next to you]. [You can avoid that inefficient trip] by sending it [locally] over WiFi."
"12% at least being offloaded in Australia by 2016 and it could be even more"
They've not been announcing changes yet because the infrastructure for offloading isn't ready or required - indeed, it's still spruiking its phenomenal-when-it-works 4G system. But Cisco maintains that even Telstra realises that future growth demands and increased 4G adoption mean the networks won't be able to cope with all that traffic without offloading it via WiFi. The amount being offloaded in 2015 might look small, but it closely resembles the entire amount of mobile traffic for 2012!
A WiFi future
Spectrum (the airwaves where wireless signals travel)
With so much talk about the growth in wireless and wireless futures it's easy to forget that WiFi is simply a way of accessing the nearby wired network, wirelessly. As Pepper puts it:-
"There is no such thing as a mobile network... the network is not mobile. I am. The network is fixed"
The theme of Cisco's report is incessant - the growth of data size, data speed and latency is so huge for all forms of broadband connection, be it wired, 3G/4G or WiFi, that ultimately everything is connected to the same fibre network with as much traffic being offloaded from the wireless networks as possible. Another important reason is because when many wireless devices communicate with networks or other close-by devices the airwaves become full and performance degrades. The more people, the more devices, the more data required, the bigger problem this is.
We've already seen how regular mobile phone networks will struggle with the massive imminent growth in mobile traffic. However, there are issues and challenges on the wider and smaller wireless scales too.
Congestion is less problematic in sparsely-populated areas like rural Australia where 'Fixed Wireless' is being applied as part of the NBN's fibre to the node policy. But Pepper says, "It's absolutely clear that there is a huge need for the mobile operators to have more spectrum for the outdoor macro cells which is why the digital TV transition is so important. That spectrum is ideal in those frequencies to go long distances. Radio is going to be great in rural areas, because there isn't going to be that contention... you can do point to point [connections]... Radio, especially at 600 -700MHz, over long distances, [has good] propagation characteristics."
Time will tell how true this is. It may be that NBN Co's imminent Fibre Extension Program, which offers people who have been assigned a fixed wireless NBN connection the ability to upgrade to a fibre connection (at 'cost' price - not for profit), may influence matters, but we'll deal with that elsewhere.
But there are potentially significant close range problems. Pepper says, "All tetherless devices... are going to overwhelm the traditional methods of connecting devices. In addition for needing spectrum for traditional networks we also need to open up more frequencies for short-range devices."
Kevin Bloch continues to describe the problem, "The more users connecting things like their televisions means you get very poor performance." Bloch points out that in January we got our first glimpse of Super High Definition "4K" televisions. These require 300Mb/s of compressed data to operate. If there are three in a house, then wireless transmission (at close range or at distance) is going to be out of the question. "These aren't the future. They're here now."
Add to this the fact that many televisions and other devices have built-in cameras and add to it the growing trend of people interacting with phones and tablets while a television show is running, and everything points to congestion at a domestic level. The brand new 802.11ac WiFi standard which is faster and more-efficient than current WiFi will help but the more spectrum there is available to spread the device signals across, the better.
Machine to machine (M2M) wireless communication is only going to increase. We already see examples of cars communicating with iPods and phones and people beaming video from their mobile devices to their televisions in the same room.
Pepper sums it up by saying, "The statistics illustrate the need for WiFi connecting to very close fibre... Symmetrical bitrates are also going to be more important."
The Korea/Japan conundrum
The report also illustrates why, despite fast infrastructure, online services had not flourished as expected in countries like South Korea and Japan. Cisco uses the World Economic Forum's The Global Information Technology Report to explain why. This immense study describes each country's physical ICT infrastructure in addition to the quality of the ICT Ecosystem in which it operates. A summary of the criteria can be seen here:-
Pepper says that "In many ways Australia is ahead of Japan and Korea. For instance, it's been a struggle to get small businesses to use broadband in Japan... It's difficult to start a business in Korea and there are taxation issues."
Regional comparison
Comparison between countries with similar per capita income.
Sweden currently comes top for both infrastructure and ecosystem. An example of its innovative broadband usage can be seen with this electric car charging program which makes use of a power distribution 'smartgrid', mobile communication and location based services.
Australia currently could do better for infrastructure but has different geographical issues compared to most other countries. The NBN should also rocket the country to the right of the graph over the coming decade. However, it may be that, in isolation, NBN-ready Tasmania could be viewed as a global leader on its own.
There is significantly more detail in the report which I won't dwell on here.
Summaries and advice
Pepper says, "All roads lead to fibre. The backhaul [cross country broadband connection] is no longer just the core of the network, it's not just the middle mile, it's now how I connect every device. It's my WiFi base station at home in the office, it's outdoors or the antennae on the mast on top of a tall building.
"Our only advice to network planners is that fibre is going to have to be pushed very very deep, whether it's to the premise or the cabinet right outside the premise. The question is, "What do you do from the node?" If the node means using existing twisted pairs... it's not going to be good enough. If the node means coax and it's close, it might be good enough. But if I'm going to have to replace the node to the home anyway, it may as well be fibre."
When asked, "What about 4G?" he said, "If it's a dedicated fixed radio link with no contention, you probably can make that work. But on a 4G mobile network that's not the case because the whole neighbourhood is going to share that. If each of us is on 4G and we are going to be consuming 28x [what we are now] and if each of us are going to be consuming multimegabit in our homes... in point to point mode, in the bush, where there are multiple flavours of 4G [it can work]. But if I'm talking about suburban Sydney, that's not going to be enough.
"From what we see based upon the data... I don't know anything about the politics of this or who said what... I'm just looking on it as someone who looks at technology, policy, trends and data and in my mind facts matter. I also sit on the UK regulator Ofcom's board and Ofcom talks about evidence-based policy making. So I just look at the trends and the data. Considering what our forecast looks at, if I'm in an urban area, a 4G connection in an area of high-density housing, is not going to be enough. It just will not scale to the same extent as fibre deep deep into the network."
"Using existing copper depends on how clean it is and how short it is and what kind of copper it is. If Coax is already there, HFC hybrid that connects to fibre - where there are eight homes off a fibre - that works." Australia's existing HFC networks don't match this standard. The point however is moot with the ACCC recently agreeing to allow NBN Co to decommission Optus' HFC network (while allowing Foxtel's to remain exclusively for television.) Neither HFC network offers over 2Mb/s upload speeds anyway.
"Globally, the dominant way that devices will be connected will be fixed WiFi. Over half the world's IP traffic will come from WiFi in 2016."
"We are going to need fibre deep deep into the network, going to need increased 4G spectrum and really going to need to address the unlicensed spectrum inside and around buildings. WiFi is much more efficient. From a network operator point of view, one of the highest costs is the immediate handoff from cell to cell on mobile network. Passing that load on to WiFi is more efficient and cheaper.
"The OpEx (operating expense) of fibre is dramatically less than copper. So it becomes a CapEx (capital expense) issue. If you can get the capital for investing in the network... once it's there, every operator in the world finds fibre dramatically reduces OpEx. So similar prices equals higher profits."
Bloch added, "The value of [Australia's] network is not the copper wires, it's the ducts they lie in." Indeed, much of the copper is old and needs replacing, the inherrant power costs to drive it are huge and it costs $1bn per year just to maintain. Fibre is the oposite of all of that.
But Bloch adds that the whole process has been, "Distorted by politics."
Politics
When Malcolm Turnbull recently told Katherine Danks at The Telegraph that "The promotion of the NBN is based on one big lie - that the benefits of super-fast broadband can only be achieved by a fibre to the premises network whereas in truth the undoubted benefits of broadband can be achieved sooner and at much less cost with alternative technologies" he was already on shaky ground. The fast, reliable, high-bandwidth, symmetrical, low-latency broadband connections that are required for the NBN's significant enhancements to Australia's health service, power distribution, education system, business, cloud requirements and social benefits all require performance and infrastructure that goes beyond fibre to the node.
With Cisco's figures now being added to the mix it further illustrates that Malcolm Turnbull's assertions are factually incorrect.
Furthermore, as this article was being published, a spat developed between Joe Hockey and Delimiter over comments made about 4G being superior to Labor's NBN. These claims have been dismissed many times before - 4G is complementary to the NBN and not a competing technology (assuming that we're not talking solely about disingenuous marketing campaigns). It couldn't hope to remotely rival fibre for all of the reasons listed above. As such, Hockey's rebuttal of Delimiter's rebuttal, which states, "Wireless technology such as 4G has the capacity to be far superior to a fixed broadband service such as Labor's NBN" is also factually incorrect. Just to stress this again, 4G and WiFi are methods of connecting to the wired/fibre network. They can't possibly hope to carry a nation's broadband infrastructure and will start to be overwhelmed in less than four years without overhauls in infrastructure. They are certainly more-convenient and popular connection methods than wired. Perhaps that is what Hockey means?
Reading between the lines, it appears that the Coalition is sticking to its famous "12 megabits is enough" stance and believes that rolling out a cheaper infrastructure which relies on existing technology, infrastructure, data sizes, speeds, latencies and symmetries is all Australia needs.
But even if Cisco's forecasts were to double in time (and it looks more likely that the 2016 predictions will actually prove conservative again) before coming to pass, any fibre to the node infrastructure is not going to get completed before it is overwhelmed. The old networking adage is true: you can't push a pumpkin through a hosepipe.
It will be interesting to see how the Coalition reacts to these figures. Hopefully they provide food for thought. Recent history suggests otherwise - there are already many questions hanging in the air. The Coalition still has not even acknowledged the NBN's benefits to health, power, education, business and society let alone explain how 12 megabits (downstream only) is enough to cater for them. Even Stephen Conroy toned down the politics recently to pose some valid observations and questions.
If the global technology sector, the Fibre To The Home Council and Cisco's report is wrong, then the Coalition needs to say why and not just casually dismiss everything. It can't just keep saying that we don't need fibre to the premises and that fibre to the node (and 4G) is enough when there is overwhelming evidence to the contrary.
Conclusion
Technology is blind to politics. In articles like this it's impossible not to mention it but valid conclusions can't be drawn, by any side, which rely on political affiliations. The technology is either valid or it isn't. As Pepper says, "Who pays for it is a separate debate. As a matter of it being technology and demand, I don't see there being any question."
The last bastion of valid discussion regarding the NBN is "Who should pay for it?" - hence the calls for a cost benefit analysis, although even this question has largely been answered. Based on all the existing evidence, the Coalition's claims regarding the technology simply don't stand up to scrutiny. If for some reason it turns out they do, then they need to explain why just about every expert on the matter has got it so wrong.
