Part 3: 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.

CVC Artificial Scarcity

It’s an all too familiar story around the country. After waiting for what seemed like an eternity for a first-world broadband service, you receive that divine moment – the NBN “Ready for Service” card in your mail box. You spare no moment as you contact your RSP and get connected to the government’s new information super highway.

But excitement rapidly turns to dissatisfaction in a couple of days when your peak night time speeds plunge to a crawl. You wished you remained on your old ADSL2+ service.

MPs (Member of Parliament) mailboxes are filling up with complaints just like this from constituents according to Senate hearings. Many want to switch back to their old ADSL2+ service, but are unable.

Some individuals incorrectly blame the “second-rate” FTTN network, but truth be told, it effects the entire NBN.

As reported in Part 1, I was a happy camper come early November 2014 when my 50/20 FTTP connection finally went live.

On March 25th 2015, movie streaming service Netflix officially launches in Australia. Much to the anguish of Rupert Murdoch, Netflix was an instant hit. Murdoch, 50 per cent owner of cable provider Foxtel, was an avid supporter of the coalition’s slower broadband network as expressed through his, somewhat biased and often factually incorrect, Australian newspapers.

Just days after the launch of NetFlix, my fibre connection started to grind to a halt. During the day, I could achieve a brisk 47 Mbps, but come 6 to 10 pm at night, I could only squeeze 6 Mbps. I fired a fault request off to my long term RSP, Internode, and waited for a response. They acknowledged there was insufficient capacity on my shared CVC – the link between the NBN POI (Point of Interconnect) and my RSP. Engineers had ordered additional capacity.

Within a week or two, I was once again a happy camper until June 2015. Again, my RSP had ran out of CVC capacity.

So what does a subscriber do if he is not getting what he pays for? He switches to a lower speed, and this is precisely what I did. By the 9th June, I had a 25/5 service. I’ve never switched back.

Even as I write this series, on Easter Monday night, my peak speed on fibre is 9.86 Mb/s.

NBN Wholesale Pricing

In simple terms, NBN’s wholesale price consists of two components – the AVC and CVC.

Each subscriber requires an Access Virtual Circuit (AVC) – think of it as a recurring port fee. A 12/1 connection costs $24 a month, 25/5 – $27, 50/20 – $34, 100/40 – $38, 250/100 – $70, 500/200 – $100 and for Senator Conroy’s fibre zealots on 1000/400 – $150.

But the AVC only gets the subscriber a port on the NBN. The RSP now needs a connection into the NBN called a Connectivity Virtual Circuit (CVC) to send internet traffic to the subscriber. This aggregate connection serves all subscribers connected to the POI.

The pricing of the CVC is considered the biggest threat to the sustainability of the NBN and has been an ongoing, controversial issue since day one.

CVC pricing was initially set by NBN at $20 a mega bit per second, per month. According to iiNet founder and now NBN board member Michael Malone, in October 2013 the price was roughly four times more than what it costs RSPs to deliver the same traffic on their own “on-net” infrastructure – that is their own DSLAMs installed in Telstra exchanges.

After a review and consultation period, an inconsequential 12.5% reduction was announced in November 2014, effective 1st February 2015. This would bring the price down to $17.50/Mbps.

Many argue the cost is a “virtual limitation” or “artificial scarcity” – most of the network cost is fixed and it costs NBN very little to offer higher speeds.

Optus Head of corporate and regulatory affairs David Epstein says “In effect, CVC pricing forces RSPs to have little or no throughput headroom in peak times, leading to adverse customer outcomes.” Epstein is of course referring to the congestion experienced by most ISPs at peak times.

According to figures presented by NBN Chief Bill Morrow during a Senate hearing in February 2016, 33 per cent of fibre subscribers are on 12/1 and 45 per cent are on 12/5. Only 16 per cent of subscribers are on 100/40. Subscribers on 1000/400, 500/200 and 250/100 range in the single to double digits.

This compares to New Zealand where 46 per cent of Chorus’s fibre customers on are on plans of at least 100Mbps.

In the same Senate hearing, Morrow reported 128 gigabytes per month was now being transferred over the NBN and as a result of the rapid increase in usage, thanks in part to NetFlix, NBN “have been evaluating a new pricing scheme for our CVC construct”. The increase in data saw ARPU (average revenue per user) rise from $39 in December 2014 to $43 December 2015, caused by a 30 per cent increase in data usage.

Earlier this month, NBN announced the overhaul of CVC pricing, introducing tiered discounts that scale with CVC bandwidth increases. The “dimension-based” pricing will start June 1st and will encourage ISPs to purchase more bandwidth.

Despite the change, RSPs still believe the charge is exorbitant. The CVC charges make up a good portion of NBN revenue enabling it to meet internal rate of return metrics.

High CVC costs can also help support the political agenda and artificially reinforce that there is little demand in higher speeds, but it comes at a big cost to innovation, business efficiency and country standing. In the most recent quarterly “state of the internet” report published by Akamai Technologies, Australia plunged to 60th rank in the world for internet speeds.

According to a recent World Economic Forum report, Australia is the lowest scoring country for affordable internet access. 57 per cent of households with incomes under $40,000 do not have internet access.

ACCC – The Anti-Competition and Consumer Commission

Australia’s broadband industry has seen rapid consolidation since the decision was made to have 121 points of interconnect in August 2012.

In a submission to the ACCC, dated the 1st February 2011, John Lindsay, General Manager Regulatory and Corporate Affairs for Internode said: “The 120 POI interconnect model will significantly increase the costs of interconnecting with the NBN which will inevitably result in smaller service providers either acquiring service from a sub wholesaler, which is financially inefficient, selling out to a competitor or exiting the market.”

On the 23rd December 2011, Internode does just that and sells out to iiNet. It followed iiNet’s acquisition of TransACT a month earlier. In the years earlier, iiNet had already acquired WestNet, NetSpace and AAPT’s consumer operations. In August 2013, iiNet also gobbled up South Australian based Adam Internet.

M2 Group purchased Primus Telecom in June 2012 and budget ISP Dodo in 2013. It also acquired Eftel, owner of the aaNet, Engin and Club Telco brands, in the same year.

By early 2014, TPG Telcom had swallowed AAPT wholesale. TPG already owned backbone fibre provider and PPC-1 submarine cable owner, Pipe Networks.

In August 2015, TPG Telecom with an estimated 748,000 subscriber base acquires iiNet with 975,000 subscribers. Number 2 telco at the time, Optus, had 975,000 subscribers. The ACCC initiates a public review under the Competition and Consumer Act 2010 that “prohibits acquisitions that would have the effect, or be likely to have the effect, of substantially lessening competition in a market” and rubber stamps the acquisition.

One month later, in September, M2 Group announce plans to merge with Vocus to create a $3.8 billion entity. The ACCC again rubber stamps the merger.

Australia, today, effectively only has four retail service providers – Telstra, TPG, Optus and M2.

With little retail competition, what incentive is there for retail service providers to invest in adequate connectivity virtual circuit capacity.

Skinny Fibre / Multi-Technology Local Fibre Network (MT-LFN)

As we learned in Part 1, the orignal FTTP NBN required above ground Fibre Distribution Hubs (FDH) to be scattered about the neighborhood. The FDH would optically split the signal and centrally service up to 576 connections. Together with thicker fibre optic cabling to support the 576 connections, expensive duct and pit augmentation was required.

What if you could have multiple smaller splitters scattered around in the pits?

This is precisely what MT-LFN or Skinny Fibre technology sets out to achieve. There are no FDHs in this architecture, eliminating the requirement to have cabinets and the civil work associated with them (duct, augmented pits and concrete foundations).

The optical splitters are now distributed among new, smaller footprint multi-ports with integral splitters, while still maintaining the same 1:32 split ratio used in the existing GPON. The distribution fibre network is joined to the local fibre network with ‘Flexibility Joint Locations’. By embracing this architecture, smaller 12 fibre ribbon cables can be used, instead of thicker 72 fibre ribbon in the local fibre network.

This new technology is touted as zero civil. The smaller cables and integrated splitters can all fit in existing pit and duct infrastructure. No extra civil works are required, saving both costs and reducing the build time by weeks.

According to leaked documents, NBN overlaid MT-LFN on a distribution area in Gosford to better understand the architecture. They tested two scenarios – a smaller FDH both above and below ground, and a cascaded split GPON based on the architecture described above. The baseline was the current local fibre network architecture with above ground FDH and was costed at $542.69 per premise. The smaller FDH scenario achieved a 36.8 per cent cost saving at $343.02 cost per premise. The cascaded split GPON came it at $217.08 per premise, a substantial 60 per cent saving.

With a very favorable outcome, a decision was made to trial MT-LFN in two FTTP areas in Victoria – 3BRA-10 (Ballarat), and 3KAL-07 (Karingal) in the second half of 2015.

Broadspectrum undertook the project development services for both 3BRA-10 and 3KAL-07, and construction for 3KAL-07 only.

In a Senate hearing on Tuesday the 15th of March 2016, Broadspectrum Project Director and Executive General Manager was questioned about the costs, but both were extremely cagey. Broadspectrum had rolled out 3KAL-01 to -06 using the current local fibre network architecture and 3KAL-07 with Skinny Fibre. Broadspectrum did acknowledge the cable was thinner and no fibre distribution hubs were used, but would not comment on the final invoice cost for 3KAL-07.

As demonstrated by the trial of 4,500 premises in 3BRA-10 and 3KAL-07, Skinny fibre is a technology available today and can substantially reduce the cost of a FTTP rollout.

Fibre to the Distribution Point (FTTdp)

Fibre to the Distribution Point (FTTdp) is exciting new technology that could be Australia’s savior and an excellent, cost effective compromise between the political gridlock of Labor FTTP vs Liberal FTTH.

Labor proponents see this technology as a logical option to a full FTTP network, bringing fibre closer to the premises and confirmation of the long term future merits of a full fibre broadband network.

Liberal proponents argue FTTdp leverages the copper network, albeit only down your driveway, and is endorsement that copper still has a significant role to play.

But the consumer, fed up with politics and the wastage, see it as a choice. They can choose, perhaps at a small cost, to have fibre to the premises or to stick with copper down the driveway.

It also offers a reasonable upgrade path in the future without having to replace large chunks of the network, as is the case with FTTN or HFC.

One of the problems experienced with the FTTP rollout is the approximately 10 per cent of the lead in conduits (LIC) between the home and the street that are blocked, requiring some civil work. According to leaked documents, the cost impact was $991 per premise if a new LIC was installed as a bulk drop at the same time PCDs were installed, or $1291 if a contractor had to return later.

FTTdp effectively moves the DSLAM (fibre to copper converter) from the street cabinet down the road to the pit out the front of the home. Each FTTdp DSLAM, called a Distribution Point Unit (DPU) is expected to support 1 to 8 ports.

The DPU then uses the existing copper pair from the lead in conduit to bring broadband into the home.

The DPU converter is “reverse powered” from the end user premises. A special plug-pack, similar to a PoE (Power over Ethernet) injector would power the equipment in your pit, eliminating the requirement for NBN to provide power to road side cabinets and the associated OPEX costs. As we have found in part 2, getting power to the nodes have been a significant challenge, adding to delays and missed targets.

The DPU would also have a bypass PSTN mode facilitating on demand transfer to the NBN, just like with the FTTN rollout, today. The existing copper line from the exchange could service the premise via the DPU. Once the subscriber applies reverse power to the line, the copper lead-in would switch over to the NBN.

The very short copper lead-in between the house and pit would support faster connections than FTTN and would be upgradable to G.fast (150Mbit/s to 1Gbit/s). As there are no shared cables, there is no need for vectoring technology.

According to leaked documents, FTTdp is about $400 more expensive than FTTN per subscriber, but mitigates many operating costs such as electricity costs, copper maintenance and future capital costs in eventually upgrading the aging copper network to fibre.

On the 18th March 2016, NBN announced a small field trial of FTTdp to 30 premises in Melbourne and Sydney consisting of NBN staff. The trial will begin in April 2016.

Call for action

Australians are being forced to switch to a monopoly wholesale network with a broken wholesale CVC model. This is the single biggest factor driving down the take up of higher speed plans, not the technology being deployed to date. High CVC pricing is leading to artificial scarcity and adverse customer outcomes.

The NBN is not just a high speed network to watch 4K movies on, competing with Murdoch’s Foxtel. It is a network for the 21st century, it will help Australia innovate, create new jobs, work smarter and improve business efficiency though access to services such as the cloud. It will enable better provision of health & education through ehealth and telehealth initiatives, in video conferencing and access to online education.

If Australia wants to drive the “ideas boom” (Turnbull’s catch phase), artificially crippling the NBN though high CVC charges and internal rates of return is counter to his strategy.

To be a good economic manager, the government must have the foresight to look at life cycle cost analysis when comparing capital and operating costs of various technologies – in what is Australia’s largest infrastructure project. Band-aiding ageing copper and HFC networks might be cheaper, ignoring operating costs, but it will eventually need to be replaced with fibre in the future. It makes sense, where possible, to build the network properly the first time round.

Skinny fibre has been proven though a NBN trial of 4,500 homes in Victoria and should be incorporated into the build immediately, despite saving face. With FTTdp looking favorable, due consideration should be taken when entering into contacts regarding the HFC network (probably too late). Once successful trials of FTTdp is completed, a decision should be made to abandon the FTTN & HFC networks in favor for FTTdp.

Turnbull’s FTTB makes excellent sense for many MDUs and should continue to be pursued.