Part 2: Let’s put politics aside. It’s time to fix the NBN.

This article is part of a three part series. Part 1 examines the merits and problems experienced from Labor’s Fibre to the Premise Network. Part 2 investigates the Coalition’s multi technology mix using a smorgasbord of technologies and the challenge of using aging networks to deliver super fast broadband into the future. Finally, in Part 3, we examine if Australia’s largest infrastructure project can be saved with a combination of Skinny Fibre and FTTdp, and if sky high Connectivity Virtual Circuit (CVC) pricing and RSP consolidation is creating a monopoly white elephant, crippling Australian business and innovation.

Part 2: “Fast. Affordable. Sooner” – April 2013

The Coalition’s broadband policy, “Fast, Affordable and Sooner” was launched in Rupert Murdoch’s Sydney-based Fox Studios on the 9th April 2013.

The modest target was for every household to have a download speed between 25 and 100 megabits/sec by late 2016, and 50 to 100 megabits/sec by 2019. The Coalition claimed the plan chosen by most users in 2021 would be only 12Mbps. (As we learned in part one, just months later on the 30th June 2013 the average speed selected on the NBN Fibre network was already 39 Mbps.)

By leveraging Australia’s existing copper network infrastructure, it was forecast this affordable broadband rollout would cost $29.5 billion (comprising of $20.4 billion in capital expenditure) and would be 33 per cent less than the current government’s $44.1 billion. Not only cheaper, it could also be completed two years sooner, by 2019.

Under a coalition government, once former OzEmail chairman, Malcolm Turnbull would be appointed as Communications Minister. The then opposition leader, Tony Abbott, invoked laughter in June 2013, when he proclaimed Turnbull virtually invented the internet in Australia:

“We have a strong and credible broadband policy because the man who has devised it, the man who will implement it virtually invented the Internet in this country. Thank you so much, Malcolm Turnbull.”

NBN Version 2: Multi-Technology Mix

On the 18th September 2013, a Liberal government was sworn in after winning the election. Shortly after, the new Communications Minister, Malcolm Turnbull, got to work asked for the resignations of the NBN Co board. By the 23rd September, the majority of the board had formally resigned.

Under the Multi-Technology Mix, a smorgasbord of technologies would be used including:

• FTTP (Fibre to the Premises)
• FTTN (Fibre to the Node)
• FTTB (Fibre to the Basement)
• HFC (Hybrid Fibre Coax)
• Fixed Wireless
• Long Term Satellite

Fibre to the Premises (FTTP) would continue to be rolled out in Greenfield sites and would eventually make up only 20 per cent of the NBN footprint.

Fibre to the Node (FTTN) and Fibre to the Basement (FTTB) would be the most prevalent technology, taking a 38 per cent share, while HFC would service the remaining 34 per cent of mainstream connections.

Unchanged from the previous government’s policy, 5 per cent would be serviced by Fixed Wireless and 3 per cent by Satellite.

Fibre to the Basement (FTTB)

One month following the change of government, on the 22nd October 2013, Malcolm Turnbull announced his first positive tweak – Fibre to the Basement.

As discussed in part 1, approximately 30 per cent of the Australian population lives in apartments or units (MDUs). Getting fibre optic cable though the core of these buildings was a challenge from both the physical roll out, and from the owner’s corporation/body corporate approval process.

The new plan would see fibre optical cabling laid to the basement or MDF (Main Distribution Frame) – the communications room of the apartment building. The building’s existing copper network would then deliver the last couple hundred meters, if that. This cabling was normally protected within the buildings core or cable risers and was considered to be in significantly better shape than incumbent’s copper access network.

The FTTB plan was to slash $800 million from the NBN costs. The owner’s corporation typically owned the cable beyond the MDF and hence Telstra’s approval was not needed.

Having lived in a 30 story apartment building in Chatswood, Sydney, it was something I personally believed was a deficiency in the previously government’s network. Walking the local neighbourhood, I was aware of the 5 – 8 story apartments that littered suburbs like Artarmon and Lane Cove.

Malcolm Turnbull was off to a positive start. But, could it last?

By the end of March 2015, NBN Co officially launches FTTB with approximately 2,000 subscribers.

Here we go again – negotiations with the incumbent 800 pound gorilla.

In the weeks prior to the election, Malcolm Turnbull expressed his views that Telstra would simply hand over their copper network for nothing, under the existing $11 billion agreement with NBN Co.

The original agreement was to utilise Telstra’s pit and ducts infrastructure, but now NBN Co needed access to Telstra’s last mile copper. The coalition also wanted NBN Co to use Telstra’s HFC network.

On the 14th December 2014, NBN Co and Telstra signed revised NBN Definitive Agreements. On the 26th June 2015, the ACCC rubber stamped the agreement.

Turnbull was right, it didn’t cost tax payers any more money. However, it cost valuable time.

Fibre to the Node (FTTN)

Fibre to the Node is the heart of the coalition’s fast, affordable and sooner multi-technology mix and the technology supporting the largest portion of subscribers – some 38 per cent of premises.

Copper broadband technologies such as ADSL, ADSL2, ADSL2+, VDSL and VDSL2 all suffer from speed degradation the further the subscriber is away from the exchange, or more precisely the DSLAM (Digital Subscriber Line Access Module) typically installed in the exchange, but which can also be found in street side cabinets and ‘top-hats’ of newer suburbs.

It is a problem that plagues current subscribers of the incumbent’s copper access network. In some areas the exchange could be as much as 7 kilometers plus away from the premise, crippling the maximum speed to a paltry couple of megabits a second, subject to copper pair diameter and condition. If you have thinner cable, a few plastic biodegradable bags, plastic PET bottles, faulty 3M gel caps or a bit of rain, then you could be out of luck.

ADSL2+, the most prevalent technology used on the incumbent’s network, by both the incumbent and competitors, has a maximum downstream speed of 24Mbps.

VDSL typically offers a maximum download rate of 55Mbps and VDSL2 increases this to 100Mbps. VDSL2 now offers an ingenious technology called vectoring that reduces crosstalk between cable pairs and helps deliver a higher bandwidth. With vectoring enabled, subscribers 400 meters or less from the node can expect to have a consistent 100/40 speed. As the length increases to 700 meters, this drops to 60/20Mbps.

If the Coalition’s plan is to have a download rate between 25 and 100 megabits by late 2016, and 50 to 100 megabits by 2019, then it would have to install tens of thousands of street side cabinets packed full with VDSL2 DSLAMs. These street side cabinets would then have to be connected to subscribers with the shortest copper runs.

And that was precisely the plan.

Few deny the fact that the incumbent’s aging copper access network is far from good shape.

Hansard from Wednesday, 23 November 1910 show transcript of the Member for Calare, Thomas Brown MP, arguing that “Copper is used [on the telegraph] for greater efficiency”. Iron had been used on the Telegraph for some 30 years and replacing it with copper was ridiculed as expensive, much like Fibre is today, one hundred and six years later.

In simple terms, the typical copper access network consists of a distribution cable containing 200 pairs from the exchange to a street side pillar. A standard pillar represents a distribution area (DA) of about 150 to 200 premises.

The pillar is a water tight, domed top metal cylinder invented in Australia and introduced into the Telecom Australia network in 1955, when Australia once led the world in Telecommunications.

Street distribution cables then connect from the pillar to the pit outside of individual homes in the distribution area. From the pit, a lead-in runs into the home or business.

In the FTTN model, a Node is placed as close as possible to the pillar. Each Node has 384 ports.

1,860 kilometers of new copper cable on order

During a Senate estimates grilling on the 20th Oct 2015, NBN Co CEO Bill Morrow declared NBN Co had purchased 1,860 kilometers of new copper cable, $14 million worth, from Sydney based Prysmian. Some commentators seized on the moment suggesting this was proof the copper network was in disrepair and needed replacing.

But the 1,860 kilometers of new copper cable is intended to be used to connect the new nodes to the pillar. Depending on the size of the distribution area, a single 100 pair or 200 pair cable will connect the new node to the pillar and subsequently the customer premise side.

To provide the transition period, allowing subscribers to migrate from the incumbent’s network, another 100 or 200 pair cable connects the new node to the pillar and back to the exchange, providing continuity of service. This cable becomes redundant in 18 months once every subscriber is migrated.

Copper Corroded Spaghetti – a 972% cost blow out.

A leaked document on the 3rd December 2015 finally provided the evidence experts were expecting. The dilapidated copper network could cost as much as $641 million to fix.

A risk assessment in the leaked document assessed the risk that the “state of copper network [was] considerably worse than expected, leading to extensive work beyond the node” and found it had a likelihood of “almost certain” and a consequence of “major”.

In strategic review number 6, one of many Turnbull commissioned hoping for a favorable outcome, the cost of copper remediation per node was put at $2,685. In the first operating plan (IOP 1.1), this figure exploded to $19,437 per node. In IOP 2.0, the cost ballooned out again to $26,115.

With 24,544 nodes, copper remediation beyond the node could now total $641 million.

The risk assessment indicated a “decision to minimise remediation during build could reduce speeds available, create additional burden on connect and hamper timely migration.”

It is currently unclear what testing or qualification of the copper network is being undertaken at build. This fundamental question was raised during the Senate Standing Committee and was taken on notice. It appears when a home install occurs, the installer will connect a “VDSL configured device” (modem) and record the Signal to Noise Ratio and Sync Rate.

For the self-install model, it is most likely no tests are actually performed on the line.

Utility Power to the Node

On the 29th February 2016, Fairfax papers had reported “NBN: Malcolm Turnbull’s ‘faster, cheaper’ roll-out falters”. It had obtained a leaked document, dated the 19th February 2016 comprehensively detailing FTTN rollout critical design metrics for the week of the 12th February 2016.

The target was to have the design completed for 1,402,909 premises, but NBN had only completed fewer than half – only 662,665 premises.

94,273 premises should have reached practical completion, and 5,943 be “Ready for Service.” Reality was only 29,005 premises had practical completion and zero were Ready for Service.

The documents highlighted the root cause for 41 per cent of the 740,244 design deficit was getting unmetered power approvals, with 91 per cent of the backlog in NSW and Victoria. In the previous government’s FTTP network, nodes are passive and do not require power, not only simplifying the network build but also reducing operating expenditure.

It is believed the 24,544 nodes planed in IOP 2.0 will pull approximately 1 kilowatt of power continuously – each. That adds up to one heck of a power bill. The original operating plan (IOP1.1) did not include power costs in the OPEX figures.

The 65,268 premise deficit for practical completion was attributed to delays in power connection (34%), power approval (25%) and material shortages (30%).

NBN Co COO Greg Adcock indicated connecting the Umina trial nodes to the Electricity network cost $6,250 each.

Fibre to the Node Commercially Launched

The Fibre to the Node product was commercially launched by NBN on the 21st September 2015.

According to the press release, NBN plans to have 500,000 FTTN premises ready for service by mid 2016 and 3.7 million by mid 2018.

Telstra & Optus HFC Networks (‘Operation Clusterfuck’)

Clusterfuck is a military term for an operation where multiple things go wrong. Operation Clusterfuck also happens to be the NBN Co internal codename for the upgrade of the Telstra and Optus HFC networks.

Under the multi technology mix, NBN Co will leverage the Telstra and Optus HFC networks built in the 1990s.

Telstra has a HFC footprint covering Sydney, Melbourne, Brisbane, Gold Coast, Perth and Adelaide.

HFC stands for Hybrid Fibre-Coax. The network, as the name suggests, consists of a fibre distribution network and coax cable with multiple amplifiers in the last mile coax segment. The module/converter between the fibre and coax networks is called a Cable Modem Termination System (CMTS).

The network is a shared medium. Each premise is connected to the same coax cable and shares the bandwidth in the segment. As more people connect to the segment and download data, the slower the network gets.

The current Telstra HFC network conforms to DOCSIS 3.0 (Data Over Cable Service Interface Specification Version 3) and supports speeds up to 100Mbps. NBN plans to upgrade to DOCSIS 3.1 in Q4 2016, providing a maximum 10Gbps downstream capability, with 1Gbps upstream.

Both Telstra and Optus currently have lead-ins to 2 million premises. According to the NBN Corporate plan, it is expected the HFC networks will serve 4 million premises or 34 percent of the population. The Telstra network is believed to be in much better condition than the Optus HFC network.

As the network is a shared medium, and as more premises are migrated onto the HFC network, it is the intention of NBN Co to undertake HFC node splitting – adding more CMTS head ends to reduce the number of premises sharing the bandwidth and improve the speed. Current CMTS can service only one segment, but with never CMTS it is possible to service two or four segments.

Current NBN design rules will see approximately 700 to 800 premises per segment and approximately 5,000 segments to service the 4 million premises in the HFC footprint. This is expected to be split to 17,000 segments in the future, reducing the number of premises per segment to approximately 235.

On the 23rd February 2015, NBN Co inked a $400 million deal with US based ARRIS to upgrade and integrate the HFC networks. The entire network would be upgraded to DOCSIS 3.0 with the option of DOCSIS 3.1 when NBN Co is ready. It is understood the first batch of 18,000 HFC NTDs are DOCSIS 3.0 are unable to be upgraded. These will be retrofitted to DOCSIS 3.1 at a later date.

Optus HFC

The Optus HFC network passes 1.7 million homes in Sydney, Melbourne and Brisbane.

On the 23rd June 2011, under the previous government, Optus signed an $800 million deal with NBN Co to progressively migrate HFC customers to the NBN and shut down its HFC network. It was approved by the ACCC on the 28th May 2012 and payment will be done in installments as customers migrate to the NBN.

Under the coalition, it was decided to use the aging network to deliver high speed broadband now and into the future. It had been hoped leveraging on the 1990’s network would increase the roll out speed and provide Australians broadband speeds, sooner.

The coalition renegotiated the earlier deal to obtain rights to use the tired network. This is known as the ‘Orion’ deal.

In a document titled “HFC Plan B: Overbuilding Optus” dated the 3rd November 2015 and leaked on the 25th November, NBN considered the Optus network not fully “fit for purpose”. It found some of the Optus equipment was “arriving at end of life” and would need replacing, segments were oversubscribed and will require splitting and existing Optus CMTS did not have sufficient capacity to support NBN services. In addition, it found degraded end-user speeds due to noise ingress into the network.

In order to provide “higher probability of success given the current state of the network” and “significant operational simplicity” NBN Co recommended overbuilding many of the 470,000 premises the network passed, migrating 240,000 subscribers to FTTP/FTTN/FTTdp and 30,000 to the Telstra HFC.

The recommended overbuilding option, the mix proposed above, would require a “significant peak funding” cost of $375 million and result in missing the HFC Ready for Service targets. Other options put forward had peak funding up to $600 million.

With NBN being a political football, some commentators were incensed the coalition had spent $800 million on an obsolete network, only to have to fork out hundreds of millions more to overbuild it.

Reality was, the previously government had already negotiated the $800 million to shut the network down, or more precisely $800 million for the migration of subscribers, not the actual network. The previous government had planned to overbuild the Optus HFC footprint with fibre. It should be noted the $800 million hasn’t yet changed hands, this will be done progressively as customers are migrated.

What it does highlight, however, is the significant challenge in cost effectively building a high speed broadband future on old, neglected and decaying infrastructure.

NBN claim no separate IT system is required to support the Optus HFC. The IT provisioning systems are HFC specific and will be used for both Telstra and Optus HFC.

TPG HFC

TPG owns HFC networks in Mildura, Ballarat, Geelong and Canberra. NBN CEO Bill Morrow has had talks with TPG but they are not interested in selling these networks. The Mildura and Ballarat networks were rolled out by regional Victoria provider Neighborhood Cable in 1996 to 2001. The assets were acquired by innovative Canberra based TransACT in 2008. In November 2011 iiNet acquired TransACT.

HFC Network Build

A three hundred premise NBN HFC pilot was undertaken in Redcliff, Queensland late last year. This has been expanded to 9,886 premises in Redcliffe, 790 in Emu Plains and 500 in Merrimac. The corporate plan states NBN will have 10,000 HFC premises ready for service by the 30th June 2016.

In 2015, just prior to Christmas, Telstra revealed it had signed a memorandum of understanding with NBN for the upgrade and expansion of its HFC network. The deal was estimated to be worth between $1.5 and $1.8 billion and would see Telstra design and project manage the upgrade. Only this week, Telstra announced the deal has been signed.

A similar deal is expected between NBN Co and Singtel-Optus, although due to the smaller reach of the Optus HFC network, it is not expected to be nearly as substantial.

Under Telstra’s very lucrative continuity deed, once NBN migrated one user to the Telstra HFC network, NBN committed to acquiring the entire Telstra HFC network. It cannot cherry-pick parts of the network.

Telstra has a contract with Foxtel for the supply of cable TV services, believed to be until 2025. NBN Co will now be responsible to Telstra for maintaining the HFC network for the provision of Foxtel until 2025. Rental payments are indexed to CPI and do not take into account the increased maintenance costs of an aging HFC network.

During the rollout, NBN cannot dispose of HFC assets transferred to it, without Telstra’s consent. After the rollout is complete, NBN cannot transfer the Telstra HFC assets to a large RSP or a third party which has credit rating below a specific limit.

In January 2016, NBN indicated Optus and Telstra are in construction trials, testing the lead-in rollout to premises passed by their network, but where the premise does not already have a lead-in. It is a construction trial only, customers are unable to connect. Special attention is being given to MDUs.

Medium to large MDUs are not well serviced by HFC. As a result, these premises as most likely to be serviced by FTTx.

NBN is also working with RSPs allowing a sandpit trial where they can connect and test their own equipment to the HFC. It is believed the NBN HFC will be one of few open wholesale HFC networks in the world.

According to a Senate Estimates hearing in February 2016, we can expect a commercial launch of the NBN HFC product in second half of 2016.

NBN Version 2 Progress

After lengthy delays in turning the ship around, commissioning a raft of inquiries/strategic reports and renegotiating contracts to purchase deteriorating copper and HFC networks, the NBN roll out is once again picking up speed. As of last week, the 7th April 2016, NBN Co have 1,969,933 premises declared “Ready for Service” and 917,350 premises activated. During the week ending 7th April, NBN Co declared another 8,674 premises Ready for Service.

On the 23rd August 2015, NBN Co announced peak funding costs have increased to $56 billion. Despite costing $56 billion, PricewaterhouseCoppers have valued a completed network at just $27 billion as a result of using a mismatch of obsolete technologies.

Comparing headline capital expenditure figures between a full FTTP roll-out and Multi Technology Mix does not take into consideration operating costs and the expected life of the asset.

When labor was confronted with needing new wheels, it brought a brand new car. Liberal on the other hand decided to upgrade the engine at considerable capital and is now faced with high maintenance costs, bald tires, a clunky gearbox and a corroded radiator.

The most recent quarterly “state of the internet” report published by Akamai Technologies found Australia fell to 60th rank in the world for average peak connection speeds, down from 30th place when the Coalition took office.

Coming up in Part 3, we examine if Australia’s largest infrastructure project can be saved with a combination of Skinny Fibre and FTTdp, and if high Connectivity Virtual Circuit (CVC) pricing and RSP consolidation is creating a monopoly white elephant, crippling business and innovation in the process.