Robin Eckermann, an adjunct professor at Canberra University and one of the people behind the TransACT network, was invited, along with a few others who have been prominent in the discussion around the network over the past several years, to offer his views about the next stage of development of the NBN. The views of telecommunications consultant Paul Budde were aired recently.
Eckermann said he was always concerned when a near-monopoly asset was put into private hands. “And the NBN is now a fixed broadband monopoly in many areas,” he said. “The government has paid to buy the customers of Telstra and Optus, and NBN Co has intentionally or otherwise inflicted substantial damage on many rival network operators. If one of the motivations behind the NBN was to break Telstra's strong grip on the Australian market, the government needs to be very careful that it does not complete a full circle and end up replacing the past situation with a bigger and uglier monopoly!”
Given that private firms were duty-bound to make returns for their investors, this would squeeze costs, including investment on network assets, or push prices higher, Eckermann said, adding that selling the NBN to a private organisation would not be compatible with maintaining it as critical national infrastructure.
To prevent this, a possible brake on a renegade future network owner was the increasing capability of mobile networks, especially with some of the advances of 5G technology. “While mobile networks will never be able to cost-effectively support the same traffic flows as fixed broadband networks, it would be easy for users without intensive capacity needs to defect from the NBN and satisfy their modest requirements on the mobile networks,” Eckermann said. "The potential for such an erosion of the customer base may help to keep the network owner ‘honest'.”
He said discussions about improving the NBN over time should also look at improving resilience – especially in regional areas that are bushfire prone.
"There were extended outages during the recent bushfires - many of them power-related – and some which could be avoided with some changes of practice. As an example, I was alerted to a vulnerability in the Durras area on the South Coast in just the past few days. Apparently power to NBN Co's fixed wireless tower runs the final few hundreds of metres aerially through a section of bush – without regular attention to hazard reduction, this could easily be a weak point at times when the need for reliable communications is at its most critical. It's not the major theme in network evolutions - but nor should it be forgotten."
The official build of the NBN was completed on Monday but there is still some way to go before tools are downed; there are many problem premises to be dealt with and those who are connected have about 18 months to go before they decide whether they want to hook up or not.
Only about two million will get the coveted fibre to their premises, while 4.7 million will have to live with fibre to the node, which means copper making up the last mile or more. A total of 1.3 million will get a shorter copper line with what is curiously known as fibre to the curb and 2.5 million will have to live precariously on hybrid fibre coaxial – Telstra's old cable network. The remainder will be on either fixed wireless or satellite
Eckermann said there was just one other means of keeping a future operator of the network in check: hefty regulation, a regime of performance obligations and pricing constraints to ensure that the NBN was able to support Australia's socio-economic growth and prosperity. “Such a regime would certainly affect the price that a private interest would be prepared to pay to acquire the network,” he added.
Asked whether the future should be all-fibre wherever possible, Eckermann said it was easy to answer yes to the first part of the query, adding that the term “wherever possible” was the problematic part. “The more one ventures into lower density areas, the less practical (financially possible) it becomes to deploy fibre all the way to the premises,” he said.
“Rather than adopting an ideological (or political) focus on technology, it would be far healthier to approach the future by seeking to understand the growth in bandwidth requirements over time, and ensure that appropriate capacity is available in a timely, cost-effective and affordable manner. Such deliberations should guide the choice of technologies - but the inevitable trend is towards deeper and deeper fibre penetration over time.”
Eckermann emphasised that the remainder of the build should be driven by “sound economic and engineering planning by suitably skilled professionals who were liberated as much as possible from political point-scoring and approached with a multi-party understanding that communications is critical to Australia's future”.
He added: “It is fair to say that I viewed some of the early efforts of NBN Co as naive and/or inept – on a grand scale. However, the organisation has matured considerably and I suspect that it is now as well-placed as any to preside over the ongoing operation, evolution and upgrade of the network.”
His view was that the NBN should be seen as a glass half-full, rather than one which was half-empty. “I believe the network should be managed for national advantage, rather than commercial gain,” he stressed. “Telecommunication networks have come to be vital infrastructure – just like roads and rail lines. I think everybody recognises the social and economic benefits that are underpinned by good connectivity. In the face of the COVID-19 pandemic, many Australians have also started to experience environmental benefits – such as the easing of pressure on roads as teleworking becomes more common.”
Eckermann said the bid to achieve average revenue per user (ARPU) of $52 at the wholesale level was driving retail prices that many Australian consumers could not pay. “It is easy to demonstrate that a significant number of Australians are already choosing speed tiers below those available to them (including on the widely-denigrated FttN technology) – and it’s hard to see past pricing as the dominant explanation,” he said. “In a simple analysis in 2018, I demonstrated that user self-choice of speed tiers was constraining Australia's ranking in broadband performance more than technology constraints – reducing the average download speed to about half its potential.”
Given that, Eckermann said he viewed the starting point for “wise management of this critical national asset being a shift towards recognising that by empowering a vigorous economy, the government would gain more in taxes, reduced unemployment costs etc than it may forfeit in wholesale network revenues”.
Reasoning from that, he said unless more of the NBN’s performance potential was made available at affordable prices, spending vast additional sums on network upgrades would benefit very few people and unlock few additional benefits. “Indeed, progress towards the target ARPU may well be negated by the defection of ‘low requirement’ users to ever-more capable mobile networks,” he added.
Eckermann suggested that one way to restructure the wholesale pricing framework would be a one-time writedown in the value of the network asset – “a reality that many past network builders have had to accept (for example, Telstra and Optus with their HFC networks, and TransACT with its FttK network). This would give new flexibility in wholesale pricing that delivered an adequate return on a reduced cost base”.
Secondly, he said, considerations about network upgrades to improve performance needed to take into account the end-to-end chain of links that connected a user to their cyber-destination. Last-mile technology (FttP, FttN, FttK, HFC, FW, Satellite) was too often the lone focus when other links, could in fact, be why performance did not meet expectations. “Other links include user devices, home wiring or Wi-Fi, modems and routers, the shared NBN connectivity virtual circuits (CVCs), backhaul between NBN points of interconnect (PoIs) and retail service provider (RSP) networks, national and international circuits, and so on,” he explained.
Eckermann said reviewing the value of the asset and relaxing the target ARPU could allow a permanent increase in CVC capacity per user so that it never acted as a bottleneck. “If and when this is done, any technology-related performance limitations on the access virtual circuit (AVC) would come into sharper focus as the target for future network investment,” he explained. “These are the two main links that are within the scope of the NBN. A third link (between the 121 NBN PoIs and RSP core networks typically located in the capital cities) was ruled out of scope in early NBN architecture decisions – and it is difficult to see that decision being overturned at this late stage.”
From the AVC angle, optical fibre all the way was the best way to obviate most practical performance limitations. “However, contrary to the views of some, it is not a digital divide where FttP is good and everything else is a disaster!” he hastened to add. “Even with FttP, the final connection between the NBN network termination unit (NTU) and the user's device is typically copper or wireless. The real issue is how close optical fibre is brought to the device that the consumer is using (with the home boundary being just one point on the spectrum).”
In this context, he said, FttP was the best AVC technology – FttK (k meaning kerb, rather than the American word curb) a close second, and FttN, HFC, Fixed Wireless and satellite a more distant third. The challenge for NBN Co would be to push fibre ever deeper into the network.
In his view, Eckermann said it would make sense for NBN Co to develop geographic "heat maps" based on the maximum speed attainable on each consumer connection, and to prioritise upgrade investment to steadily improve the weakest areas wherever they occurred.
“In the case of FttN areas, I expect the approach would be to split large FttN serving areas into smaller FttK cells and equip each cell with its own kerbside node to reduce the length of the residual copper segment,” he said. “With the latest VDSL technology, this would support performance into the hundreds of Mbps - and for users who want more (and are willing to pay for it!), the upgrade of individual AVCs from FttK to FttP technology should involve comparatively predictable and modest costs.”
In HFC areas, the approach would be similar, with smaller cells reducing congestion on the shared coax segment during busy periods.
Asked about how much would have to be spent for an upgrade that would stand the test of time, Eckermann took a cautious approach.
“The cost of upgrades will depend on many factors – but most significantly, the depth to which fibre is taken in the network,” he said.
“Physical access and the civil works involved are one of the major cost determinants. For example, where existing conduits (or poles) are in a fit state to support new optical fibre cabling, the costs will be dramatically lower than if new conduits need to be retrofitted. Of course, if physical access incurs rental figures, the ongoing operational cost burden needs to be weighed against the upfront costs of constructing new access routes.”
He said the cost of taking new cabling into homes was also very high – and use of existing copper or coax "tails" into the home represented a typical saving of more than $1000 per premise. Thus, if the aim was to upgrade all FttN/FttK and HFC areas to FttP, the cost would be significantly higher than if existing tails could be used.
“Most estimates I have seen are theoretical, often based on wishful thinking or highly selective cost figures taken from very different contexts,” Eckermann said. “Speaking from the perspective of someone who has been involved in a significant network build (TransACT's 65,000 premise brownfield FttK/VDSL build starting some 20 years ago), I would simply emphasise that the unexpected complications that invariably arise can be very expensive and push average costs upward.
“In 2009 when the NBN was first announced, my view was that the original plan (FttP to 93% of homes) would have cost about $100 billion when central costs, start-up losses and the high costs of passing some 12 million premises and connecting (indicatively) 75% of them was taken into account. The initial figure of $43 billion was never credible to anyone with real-world experience.”
Despite having been at the coalface, he said he was still hesitant to venture an estimate without actually examining the typical scenarios that would be involved.
“Some cost elements are now behind us – for example, most of the migration fees paid to Telstra and Optus. However, pushing fibre deeper into the network will involve significant civil works – and the magnitude of the undertaking will be determined by the end objective. If the end objective for the seven million (approximately) homes in FttN and HFC areas was an upgrade to FttP, I would expect the cost to be indicatively $10 billion higher than if the end objective was FttK for the same areas.”