NBN Co's great gamble

The great Hybrid-Fibre Coax gamble is well underway at NBN Co and what this will mean for consumers is discussed in Business Spectator. The promises of future upgrades and black spot remediation are under the microscope as there is increasing evidence that the short term promise of a quick boost in NBN customer numbers provided by existing HFC customers is too big a cherry for the government and the communications minister Malcolm Turnbull to overlook.

Read the full article below

NBN Co has just got a brand-new HFC division up and running which will run by chief technology and security officer Dennis Steiger. The former Shaw Communications executive has plenty of experience under his belt but navigating the murky waters of the NBN won’t be easy.

Steiger, like most of the NBN Co executives, will be well aware that the clock is ticking and failure to produce results before the next election could bring about a swift exit from a company that has experienced considerable staff turnover since the last election.

However, he will be confident that he can provide the government with a significant boost in the lead up to the next federal election by integrating the Telstra and Optus HFC networks that pass more than three million premises into the National Broadband Network .

NBN Co has been quietly building the momentum needed to complete the HFC integration before the end of 2016 and there does not appear to be any reason why this should not occur.

For NBN Co the purchase of iiNet’s TransACT Fibre to the Premises (FTTP) network in May 2013 provided a small (13,000 premises passed, planned or under construction) but that was a significant boost to connection numbers at a time when progress was urgently needed before the September election. 

A similar but significantly larger outcome provided by adding the HFC networks to the NBN would provide the Coalition government a much needed boost at a time when the opposition parties will be reminding everyone of the government’s unachievable promises made in the lead up to the previous election.

The DOCSIS 3.1 upgrade

To counter the critics the government is likely to announce an upgrade program for NBN Co’s HFC network from DOCSIS 3.0 to 3.1 when it becomes available for residential customer deployment in 2016.

DOCSIS 3.0 can be used to connect customers at about 1 Gbps downstream and 216 Mbps upstream but due to infrastructure cost, perceptions of customer demand and other factors DOCSIS 3.0 in Australia is typically used to provide either up to 30/1 Mbps or up to 100/2.4 Mbps.

And it should be remembered that HFC connection speeds are only “up to” speeds with no performance guarantees provided due to the hybrid cable that utilises fibre and copper segments.

DOCSIS 3.1 is planned to provide customers with gigabit connections with significantly improved upload speeds than what are commonly available with DOCSIS 3.0 networks. But it should be remembered that the NBN Fibre-to-the-Premises (FTTP) network currently offers 100/40 Mbps or the faster 1000/400 Mbps and by 2018 NBN Co will be able to upgrade the FTTP network to provide the faster 10/4 Gbps.

iiNet recently upgraded its HFC offering in Geelong, Ballarat and Mildura where it passes about 80 per cent of premises to “NBN like” connection speeds of up to 100/8 Mbps ($69.95 per month), 50/8 Mbps ($64.95 per month), 25/5 Mbps ($54.95 per month) or 12/1 Mbps ($49.95 per month).

The prices are very competitive, especially when the bundled extras are included, but the internet service provider’s HFC offering still falls well short when upload speeds are compared. For example iiNet’s fastest HFC connection provides up to 100/8 Mbps and NBN Co’s equivalent FTTP offering provides 100/40 Mbps, not “up to” 100/40 Mbps.

NBN Co will seek to provide customers with equivalent connection speeds where possible irrespective of the technology into customer premises. What this means is that NBN Co’s HFC network will need to be upgraded to provide up to 100/40 Mbps.

This can be done with DOCSIS 3.0 but there is a catch. Customers utilising HFC share access to the cables and DOCSIS 3.0 was designed primarily for fast downloads which means that there is far less capacity available for uploads.

Capacity management is crucial

One positive factor of the NBN FTTP network is that it is new and designed for low congestion by limiting the number of customer connections per fibre to 24, with the capability to increase to 32 if additional premises are built (for example units built to replace a house). The number of customers on a HFC cable run is already far too high for upload speeds above 8 Mbps to be offered without significant congestion occurring.

Telstra and Optus do not provide their HFC network design rules and it is still unclear whether NBN Co will release updated network design rules early in 2015, as suggested by NBN Co chairman Dr Ziggy Switkowski during his talk at the Committee for Economic Development of Australia business luncheon held in September. 

HFC network design takes into account many factors, including the number of customer connections to cables, connection speeds, congestion and so on. A US cable expert Peter Jacoby told  ArsTechnica  in December 2013 that “if customers aren't getting their promised speeds, ‘we need to upgrade or split that group’ so that fewer people are served by the same node." 

"This capacity management is the key; it's the secret sauce to offering a service that ensures customers get what they want, because the customers are always using more [than they did last year].’"

The Telstra and Optus’ HFC makeover won’t be easy or cheap

What this means is that NBN Co will have a major task on its hands to upgrade the Telstra and Optus HFC networks in order to provide customers with up to 100/40 Mbps connections that are not heavily congested during peak periods.

It is estimated that NBN Co’s HFC networks will only pass about 80 per cent of customers in the HFC designated areas initially but the exact coverage won’t be known until NBN Co gets full access to the Telstra and Optus HFC networks.

NBN Co may decide that provision of up to 100/40 Mbps over the HFC network will not be possible without extensive remediation and the DOCSIS 3.1 upgrade forcing a delay of this offering into 2017/2018.

In any event the DOCSIS 3.1 upgrade will be expensive, probably around $1 billion and there will be an additional cost of $2 to $3 billion to reduce the number of customer connections on each cable to reduce congestion.

There's also the need to run cable to the premises missed by Telstra and Optus including apartment buildings within the HFC areas though these could be added to the NBN utilising Fibre to the Basement and VDSL2 with vectoring. However, the cost of doing this in a HFC area should be included in the total cost of providing the NBN in that area, in other words bundled with the cost of providing the HFC.

The up-front cost of the NBN Co purchase of the Telstra and Optus HFC networks lies somewhere between $2.5 and $4 billion in 2012 dollars when you take into account the $800 million to be paid to Optus and the complex charges paid to Telstra for every customer moved to the NBN; and the infrastructure leasing, maintenance and operational charges buried in the agreement with NBN Co and myriad other agreements.

The government is likely to resist efforts for the HFC network cost over the period 2012 to 2030 to become public but it will be a sizeable amount of the total capital expenditure (CAPEX) and operational expenditure (OPEX) spent by NBN Co. Without transparency it will be difficult to identify if NBN Co has actually saved anything by utilising HFC other than time.

There is no doubt that HFC integration into the NBN is a good move in the short term as a quick fix whilst FTTP is rolled out but a difficult decision will need to be made before funds are spent to upgrade the HFC network about whether it is viable or whether FTTP should be overbuilt.

Do you think the NBN Co HFC gamble will be value for money over the next 20 to 30 years?

Mark Gregory is a Senior Lecturer in the School of Electrical and Computer Engineering at RMIT University.