LTSS Base Station: Sunset

NBN lets down Tasmania, again.

Tasmanian communities of Queenstown, Rosebery, Zeehan and Strahan will be forced onto an already “severely oversubscribed” satellite beam.

The company building the National Broadband Network (nbn) has revealed that they will no longer provide a fixed-line or fixed wireless solution to major settlements located along the west coast of Tasmania.

Up till July this year, nbn has released various rollout plans showing the towns of Queenstown, Rosebery, Zeehan and Strahan as candidates for the Multi-Technology Mix (MTM) and Fixed Wireless rollouts which would have seen Fibre to the Node be rolled out to the majority of premises.  The area was also expected to receive a Fibre to the Premises rollout under the previous all-fibre NBN policy.

  • Queenstown: 1,300 FTTN
  • Rosebery: 600 FTTN, 300 FW
  • Zeehan: 500 FTTN
  • Strahan: (originally planned for NBN Fixed Wireless)

However, when questioned by Senator Anne Urquhart at a Senate Estimates hearing last month — nbn‘s chief executive Bill Morrow revealed that they have re-allocated premises in those areas to be serviced by the recently launched NBN Satellite.  Citing high costs of up to $20 million to build out a second fibre path to the west coast towns, the executive said:

Because of the cost to provide fibre services in the backhaul sense to serve within the FTTN. The area only has one fibre path going out to it, and you need to have two for redundant based services. The cost—and we have looked at it a number of different times—to provide fibre out there in a different path makes it exorbitantly expensive.

The move to satellite has also been independently confirmed using NBN’s internal technology modelling.  However, the towns affected are all currently being serviced by Telstra ADSL/ADSL2+ services, with Queenstown having access to the Telstra 4G network as well.

“Severely Oversubscribed” Satellite Beams

The beams servicing this area, number 54 and 56, has also been identified by NBN’s Fixed Wireless/Satellite Strategic Review as being “severely oversubscribed” prior to the redesignation of the fixed-line footprint in July.  Adding another 3,000 premises would not improve the satellite congestion anticipated in those areas.

NBN Long Term Satellite beams
Diagram showing NBN Co satellite beams. Red beams are severely oversubscribed, yellow is oversubscribed. (FWSat SR, 2014)

However, given the availability of existing ADSL/ADSL2+ services in those towns, it is unlikely that residents will switch to the National Broadband Network due to increased latency.  As revealed last monthnbn will implement a fair use policy for the NBN satellite.  While the final policy is yet to be confirmed, the first version released saw a standard quota of 75GB per month… a far cry from the current data quotas on comparable ADSL plans.

Microwave backhaul

A report written by Engineers Australia in 2010 stated that the affected towns (amongst others) were serviced by Telstra using microwave backhaul links rather than fibre at the time it was written.  However, the testimony given by nbn executives at Senate Estimates suggests that a single non-redundant fibre path has been built since the report was written.

Originally, it appears that nbn had planned to build its transit network out to Queenstown (see diagram below, published in March 2014) using a single non-redundant spur fibre path from Sheffield or Burnie.

NBN Co's transit network
NBN Co’s transit network as at 31st March 2014


$20 million to build a redundant path to service ~3,000 potential customers does seem unreasonably high.  I don’t think it’s wise to go ahead to do spend that money.

However, it does beg the question why NBN cannot use their own microwave links as the redundancy path to service the west coast communities.  Given there is supposedly already a single fibre path that nbn can utilise, using microwave links as a redundancy path would surely be cheaper than fibre — right?

nbn has effectively neglected these communities. I doubt anyone who lives in an area with existing, well-established communications infrastructure like ADSL/ADSL2+ connections and Telstra 4G mobile reception would opt for a NBN Satellite connection given their smaller data allowances (compared with fixed-line DSL) and higher latency.  This is most unfortunate, given National Broadband Network is supposed to fix and improve connectivity around Australia — not offer a degraded version of it.

I’ve seen isolated cases like this in the past — people able to access existing DSL broadband but placed on the satellite… but not to this scale.  Not entire communities like this.

Yes, nbn‘s current Government policy is to build the network out “at the least possible cost”… but that doesn’t mean putting   thousands of premises into an already severely congested satellite beam!  Our former communications minister, now PM’s buzzword of being “agile” seems to be lost at nbn.  Surely as a special case, there can be alternative arrangements made for the redundant path?

Read More

Fibre networking at an NBN Point of Interconnect

M2’s wholesale nbn could increase aggregation competition

The introduction of M2’s nbn wholesale offering could spark increased competition in Layer 2 and 3 wholesale broadband market

This week, the M2 Group announced the launch of their new “NBN Connect” offering for Layer 2 and Layer 3 providers.  “NBN Connect” will provide three core products to resellers and layer 2 service providers, helping them connect to all 121 NBN Points of Interconnect: “Brand Connect” which is essentially a whitelabel service, “Reseller Connect” an end-to-end Layer 3 network offering and “Network Connect” — a Layer 2 aggregation offering. This comes as nbn appears to be developing an AVC trunking product to help smaller service providers reach more points of interconnects.

M2’s offering competes directly with AAPT’s National Wholesale Broadband product, who also provides Layer 2 or Layer 3 services to retailers over the NBN.  After the closure of Nextgen Network’s NBN “Virtual Connect” offering around twelve months ago, AAPT became appeared to become the Layer 3 provider of choice with many virtual service providers.

M2 could also have a competitive edge with its wholesale offering with its “Brand Connect” product — a bespoke solution which could help manage billing, provisioning and customer service for to end-user customers.  This could be a dealbreaker for customers like major supermarket chains who may not necessarily want to run their own support staff.

One thing’s for sure — increased competition in this space is a welcome sight for virtual ISPs who presumably run on razor thin margins anyway.  The planned M2 and Vocus merger, which recently received ACCC approval, could translate to better performance in the currently debatable quality of the M2 network. Hopefully this translates to greater savings for consumers in the long run.

Read More


Revenue targets: why NBN should be doing more to promote higher speeds

By actively discouraging users and downplaying the need for high speeds, nbn’s media strategy is killing its own business case

(opinion) The national broadband network is all about bringing ubiquitous, high-speed broadband to all of Australia — regardless of where you live.

However, with the shift to the multi-technology mix and in the introduction of the Fibre to the Node technology into the network — nbn is doing its best to downplay the importance or the need for speeds beyond 25 Mbps download in order to justify its technology-of-choice.

Yet, as the company reveals itself, it will rely heavily on these downplayed high-speed users 5 years time in order to meet the required revenue targets to make their Multi-Technology Mix work.

nbn’s own figures: 100/40 Mbps expected to be most common by FY2021

By the end of the rollout, nbn expects more customers will take up the 100/40 Mbps speed tier than any other speed tier on the national broadband network. (nbn has only provided FY2021 figures here, but end of rollout is expected to be the end of CY2020)

According to their own figures provided at Senate Estimates this year, 30% of all fixed-line customers are expected to take up the 100/40 Mbps tier followed closely by 29% taking up the lowest tier — 12/1 Mbps.

(N.B The following calculations assume a 100% take-up.  While the latest corporate plan only assumes a 73% take-up, magnitude (and weighted %) should not be materially affected.)

(FY2021) FTTN FTTP HFC Aggregate
% takeup Approx. premises % takeup Approx. premises % takeup Approx. premises Approx. premises Weighted %
12/1 Mbps 31% 1,395,000 25% 600,000 28% 1,120,000 3,115,000 29%
25/5 Mbps 26% 1,170,000 27% 648,000 28% 1,120,000 2,938,000 27%
25/10 Mbps 10% 450,000 4% 96,000 2% 80,000 626,000 6%
50/20 Mbps 11% 495,000 7% 168,000 5% 200,000 863,000 8%
100/40 Mbps 22% 990,000 35% 840,000 35% 1,400,000 3,230,000 30%
250/100+ Mbps 0% 0 2% 48,000 2% 80,000 128,000 1%

Source: nbn AVC profile – Question on Notice 118 (Senate Budget Estimates, May 2015)

100/40 Mbps accounts for most revenue by FY2021

But if you look at revenue figures, in user access (AVC + UNI) revenue alone, the 100/40 Mbps users account for $1.47 billion dollars per year.  At an average 1:80 contention ratio, it could amount to a total of $2.32 billion in revenue if current CVC costs of $17.50 per Mbps is retained.

The expected revenue in the 100/40 Mbps tier is expected to be double of the next highest tier in revenue terms — the 25/5 Mbps.

(FY2021) Approx premises (millions) AVC per month AVC revenue (annual, bn) CVC 1:80 revenue (annual, bn) Annual revenue (AVC + CVC, bn)
12/1 Mbps 3.12 $24.00 $0.90 $0.10 $1.00
25/5 Mbps 2.94 $27.00 $0.95 $0.19 $1.14
25/10 Mbps 0.63 $30.00 $0.23 $0.04 $0.27
50/20 Mbps 0.86 $34.00 $0.35 $0.11 $0.47
100/40 Mbps 3.23 $38.00 $1.47 $0.85 $2.32
250/100+ Mbps 0.13 $70.00 $0.11 $0.08 $0.19

Source: AVC and CVC pricing based on nbn’s price list released on 2nd November 2015

While magnitude of pricing may be altered due to pricing changes and lower overall take-up, assuming take up profiles are met — the importance of promoting higher speed plans cannot be understated.

If nbn knows what’s good for them, they shouldn’t be focusing energy on discrediting the need for speeds beyond 25/5 Mbps.  Those are simply short-term political defences which could eventually harm the revenue capabilities of the company in the long term.

It is more important than ever to promote the higher speed tiers to customers who can access the speeds, given the MTM rollout by nature will prevent some customers who want the higher speeds from getting it (I’m looking at you, FTTN).

Beyond 100/40 Mbps

Unlike the 2012 Corporate Plan which predicted ~10% of customers will take up a 250/100 Mbps service by 2020 — nbn is now predicting that by fiscal year 2021, the take up of services above 100/40 Mbps will only be at 2% for both FTTP and HFC (0% for FTTN) — or in real premises figure, around 128,000 homes or businesses.  This amounts to roughly $190 million dollars in annual revenue.

This is a substantial downgrade in forecasts and it seems the only explanation that nbn is giving is that there is not current demand for speeds above 100/40 Mbps:

“… the services that we sell, 80 per cent is 25Mbps or less, yet we offer up to a gigabit per second … this is an indic­ation, right, of what people actually are willing to pay for and what they really need.”

Bill Morrow, June 2015

Yes, the company is using current take-up information to model demand in 2020.  It seems nbn might be having some trouble understanding the broadband demand and the market in Australia.

Are executives aware that there are currently no service provider who provides speeds above 100 Mbps on a residential connection? (There are 100/100 Mbps symmetrical plans using the 250/100 tier)

There is a few reasons behind this, but firstly, nbn‘s own CVC pricing is the key inhibitor in enabling these services.  Saying “market is speaking in Australia” but not recognising they are they key inhibitors in the market is a bit of a self-fulfilling prophecy — right?

You might as well jack up all nbn access costs to a million dollars per user and say: “The market is speaking! By 2020, no Australians will want any broadband from nbn!“.  It will most undoubtedly be a true statement if the company decided to go with it.

We also need to understand the extreme costs of buying backhaul to the nbn Points of Interconnect ruled by incumbent monopolies.  Plus, there’s the high cost of IP transit in Australia (that’s the cost of connecting to the actual Internet) compared with places like the US or UK because of the large bodies of water that separate us from the rest of the world.

However, market competition means that the latter of the two (backhaul and IP transit) will likely fall as demand increases.  But CVC costs?  That’s entirely controlled by the nbn as it’s not a competitive market.

So whether or not Australia will want gigabit speeds is almost entirely dependent on nbn, and judging by the tone used by this management, it seems they neither want to promote the speed nor get the revenue — which is a pretty stupid strategy in my opinion.

Concluding thoughts

So… dear nbn,

Stop dissing users who demand high speeds.  As a taxpayer who’s tax dollars are being used to fund the project, I want it to succeed in the long term — not just politically in the short-term.

According to your own numbers, the take-up of higher speed services is paramount in ensuring the return of investment to the Government and in turn, the taxpayer.

Saying there is no demand for speeds beyond 100/40 Mbps is another one of those dangerous, broad statements that reflect the short-sightedness of the company.

When telcos worldwide are spruiking to their customers about the wonders of gigabit connections, we are special in Australia.  We are somehow unworthy, and “unneedy” of these higher speeds.

Let’s play common sense, and promote things like any sane, commercial company would.  After all, we all just want this investment to succeed.

Read More

LTSS Base Station: Sunset

Just how fair is nbn’s proposed Satellite Fair Use Policy?

Simple maths shows 2x more per-user capacity compared with interim satellite, but only 0.7x more data allowance is proposed.

After revealing the proposed restrictions and fair use policies to help manage traffic and congestion on nbn Satellite service, some residents have taken to Whirlpool Broadband Forums to express their concerns.  The primary concern being that a 75 GB 4-week rolling average limit doesn’t translate to 75 GB of “anytime data”.

According to the draft Fair Use Policy, nbn will — in addition to limiting the maximum usage per service to 75 GB on the standard CVC Class 0 — also limits the ISP-wide 4 week rolling average download and upload during peak periods (from 7am to 1am the following day, local time).  Currently, the base CVC Class 0 is proposed to be 15 GB download and 3 GB upload.

This means, if the Fair Use Policy isn’t altered before the final satellite product release, a service provider would have to either carefully balance a number of lower capacity plans with higher capacity plans to maintain an average of 15GB, or offer one plan that provides close to 15GB with around 40GB of off-peak data up for grabs for all users.

(25/5 Mbps)*
User data cap
(4-week rolling avg)
Peak Period Download
(ISP 4-wk average)
Peak Period Upload
(ISP 4-wk average)
0 $27.00 75 GB 15 GB 3 GB
1 $45.00 100 GB 20 GB 4 GB
2 $67.00 150 GB 25 GB 5 GB

Realistically, however, we can expect RSPs to oversell slightly given the 15 GB limit is averaged across the ISP using the assumption that a portion of their users will not reach the 15 GB limit.

Comparing increased capacity with Fair Use Policy

But just how fair is this 15 GB / 3 GB averaged limit being proposed by nbn?  Let’s see how it compares with the current nbn Interim Satellite Service and nbn Fixed Wireless network.

Per-user allocated capacity

Firstly, let’s check how much satellite capacity is expected to be assigned to an end user.  Since nbn imposes minimum customers-to-CVC (AVC-to-CVC) ratio (so RSPs can’t hog virtual capacity), we can calculate these numbers by dividing the CVC capacity by the maximum and minimum customers.

Estimating the per-user allocated capacity on the draft NBN satellite service FUP
Estimating the per-user allocated capacity on the draft NBN satellite service FUP

As seen above, this averages out to be around 125 kbps per user on CVC Class 0, 145 kbps per user on CVC Class 1 and 190 kbps per user on CVC Class 1.

CVC Class Maximum CVC (Mbps) Average per-user capacity allocation (kbps)
0 475 125
1 550 145
2 735 190

In comparison, the dimensioning of nbn‘s Fixed Wireless network allocates roughly 500 kbps per-user and 40 kbps on the current nbn Interim Satellite Service (ISS).  In light of this, let’s compare the percentage difference in capacity on the various CVC Classes compared to Fixed Wireless and ISS:

CVC Class Per-user capacity (kbps) % of ISS capacity (40 kbps) % of FW capacity (500 kbps)
0 125 313% 25%
1 145 363% 29%
2 190 475% 38%

So, the standard CVC Class 0 product has roughly 2 times (or is 313% of) per-user allocated capacity on the Interim Satellite Service.  It is also roughly a quarter of the capacity on the NBN Fixed Wireless network.  How do the Fair Use Policies for the ISS and Fixed Wireless compare numerically to the proposed Long Term Satellite?

Fair use policies

Interim Satellite Service: nbn‘s fair use policy for the Interim Satellite Service is structured similarly to the proposed Long Term Satellite policy.  Currently, the cap at a per-AVC (or customer) basis is 50GB.  313% of that would bring it to ~150GB, rather than the 75GB proposed.

But as we have said, this limit means little given the bulk of downloads generally occur during peak time.  While nbn doesn’t have a peak time average limit on the ISS, the company has mandated an ISP-wide averaged 9.7 GB download limit on a 4-week rolling average basis.  For this model, let’s assume that ~90% of traffic occurs during peak time — coming to an estimated ISP-wide averaged 8.7 GB download limit on a 4-week rolling average basis.  313% of that would bring us to ~27.6 GB, compared to the 15 GB proposed.

In summary, while per-user assigned capacity has increased by around 2x compared with the ISS — data limits in the fair use policy have only increased by a mere 0.72x.

Fixed Wireless: nbn‘s fair use policy for the Fixed Wireless network is straightforward.  RSPs must maintain, on average across their user base, less than 200GB download (let’s consider download only).  Like the ISS, there are no limits on peak/off-peak usage… but let’s assume that 90% of traffic is carried during peak time (meaning ~180 GB download).  If we take 25% of the Fixed Wireless fair use policy, we get 45 GB per month of peak time data per month.

Having calculated that, however, it’s important to remember that Fixed Wireless is a fundamentally different product.  Given it has far more capacity compared with any satellite product, there is more room to wriggle when it comes to the fair use policy limits.  However, for completeness sake, while the per-user capacity on the LTS is around 25% of Fixed Wireless — the data limits in the fair use policy are a mere 8.3%.

CVC Class % of ISS capacity % of ISS FUP (1) % of ISS FUP (2) % of FW capacity % of FW FUP (1)
0 313% 172% 150% 25% 8%
1 363% 230% 200% 29% 11%
2 475% 245% 300% 38% 17%

Table: % of per-user allocated capacity when compared with ISS capacity and FW capacity, and compares this with the % of the various data allowances set in proposed Fair Use Policy
(1) refers to the ISP-wide 4-week rolling average, (2) refers to the per-AVC cap (50 GB on ISS, not applicable for FW)

The summary

nbn appears to be very conservative when it comes to allocating data allowances on the Fair Use Policy.  Comparing to the Interim Satellite Service, per-user capacity has increased by between 2x to ~3x (depending on which CVC Class you choose).  However, ISP-wide peak-time data allowance only increases by between ~0.72x to ~2.5x (again, depending on CVC Class being compared).

As the writer doing the analysis, I believe there is scope for nbn to increase limits currently proposed in the draft Fair Use Policy.  Ideally, if service providers can buy two times more CVC capacity compared with what they can do on the ISS — they should be able to offer customers around the same amount more data… but I’m not going to sit here and pretend that using capacity can be calculated using simple maths.

Real networks are far more complex beasts, and simple calculations do not necessarily reflect how real networks behave.  But I believe a simple scale projection, as I’ve done above, would aid the public gaining some perspective on how little data caps have increased in comparison to the new satellite’s extra capacity.

All is not lost:

Some service providers who are helping nbn test their satellite systems before the Q2 2016 launch are actively working with nbn to get the best Fair Use Policy possible.  In the words of Paul Rees, the managing director at SkyMesh:

“We’re doing all we can to get the best FUP for customers that provides decent Data Allowances that don’t wreck LTSS performance at peak times.”

“The FUP hasn’t been set in stone yet, those numbers are just part of a draft document. So let’s have some calm, and save the ‘going nuts’ until the FUP has been formally decided.”

Wise words, Paul.  We wish you all the best in your pursuit.

Read More

NBN technician splicing fibre in a van

NBN consults on Fibre TV in new developments

New product to transmit terrestrial TV signals over NBN fibre is designed to compete with other fibre providers for new developments

The company responsible for building the National Broadband Network, nbn, has released a briefing paper on the so-called Fibre TV concept, requesting for feedback and expression of interest.  Fibre TV would take advantage of the Radio Frequency over Glass (RFoG) standard, enabling Radio Frequency (RF) signals to be transmitted over fibre.

nbn‘s new product is seen as a response to the Federal Government’s updated Telecommunications in New Development (TIND) policy which has seen increased competition for delivering telecommunications infrastructure to new developments.  Many new developments who are serviced by other fibre providers take advantage of this technology, which enables developments access free-to-air TV with antenna-less rooftops and also potentially gain access to premium channels not otherwise transmitted over free to air TV.

Previously, nbn had removed RFoG functionality from their network when it acquired the TransACT Fibre to the Premises network around Canberra.  TransACT previously took advantage of this technology, however, as part of the transition to NBN Fibre — antennas were installed in place service.

In the briefing paper, the company outlines three main steps taken to enable Fibre TV in a new development:

Firstly, a number of TV service providers will sign an “RF Light Path Agreement” enabling them to use their RFoG equipment on the nbn network where needed. These providers will typically provide free-to-air TV services, but can also provide premium services.

New developments would then, on engaging with nbn to build out their fibre network, request the company to enable this technology in the network.  The company would install an RF converter, separate to the Fibre Network Termination Device (F-NTD), to connect to the coaxial TV cable within the home.

The development would could then engage with nbn-certified TV service providers to provide their residents with television services over fibre.

Under the proposal, nbn will only make this service available to new developments greater than 250 Multi-Dwelling Units or 500 Single-Dwelling Units in the FTTP footprint.

Feedback and expression of interest in the Fibre TV service is due back to nbn by 10th November 2015.

Read More

Fibre networking at an NBN Point of Interconnect

NBN working on AVC trunking project to overcome 121 POIs

For a small fee, nbn could allow small service providers tap into its transit network to drive up backhaul market competition

nbn, the company responsible for building the National Broadband Network, has reportedly been working with smaller service providers in developing a so-called “AVC trunking” service.

Currently, service providers are required to connect their network with nbn‘s 121 points of interconnect located around Australia in order to service all of Australia.  This puts smaller service providers who do not have existing backhaul networks at a big disadvantage.

This AVC trunking project aims to allow smaller Tier 2 service providers to take advantage of nbn‘s inter-POI (point of interconnect) transit network by paying a small fee to terminate Access Virtual Circuits (AVCs) from smaller or more remote Points of Interconnect to larger depots located in capital cities.

ACCC seeks feedback from small providers

The Australian Competition and Consumer Commission (ACCC) has interviewed a number of small service providers to obtain feedback regarding nbn‘s AVC trunking project.  It’s understood that feedback from small service providers has been overwhelmingly positive.

Paul Rees, Managing Director of the ISP SkyMesh, has noted on Whirlpool Broadband Forums that: “Unless the ACCC approves this scheme, or the larger providers start offering backhaul at realistic prices, we’ll never make it to Tasmania, the Northern Territory or far north Queensland. If the ACCC is serious about maintaining competition on the nbn™ network, they will approve [CVC Trunking].”

A consolidated backhaul market

When the NBN was first established, NBN Co had preferred a 7+7 Point of Interconnect model where service providers would connect to major interconnection points in capital cities to service an entire state.  However, after lobbying from major backhaul monopoly providers from the likes of Telstra and Optus — the ACCC favoured a dispersed 121 Point of Interconnection model.

Since the ACCC decision, major acquisitions by TPG and M2 have resulted in significantly reduced competition in the backhaul market.  Most notably, TPG’s acquisition of PIPE Networks and AAPT, plus the recent Vocus-M2 merger sees almost all major backhaul providers aligning with a company with a consumer retail front.  This could allow the companies to increase wholesale backhaul costs to their competitors to lock out retail competition.

By opening up nbn‘s inter POI transit network to the AVC trunking project, it could drive backhaul competition especially to regions with less transit competition such as Tasmania and Northern Territory.  The implementation of any such project would be subject to approval by the ACCC.

Read More

LTSS Base Station: Sunset

NBN Long Term Satellite restrictions: revealed

nbn proposes standard 75GB data cap per month with 100GB and 150GB premium options to ensure satellites aren’t congested.

On Friday, nbn, the company responsible for building the National Broadband Network, released draft documents relating to the NBN Co Satellite Network set to launch in the second quarter of next year as part of the Satellite Trial Product Agreement.  The documents (see bottom of page) reveal a number of measures and restrictions expected to be taken by the company to prevent excessive use of data by end users.

CVC Classes

nbn will be introducing a concept called CVC Classes, which are essentially different levels of service on the NBN Co Satellite Network.  There are three classes, starting with the standard CVC Class 0, plus two additional premium CVC Classes — 1 and 2.  These classes are costed differently, have a different maximum contention ratio and have a different maximum data usage cap.


Cost-wise, the company will charge service providers an extra $18.00 and $40.00 ex. GST per Access Virtual Circuit (essentially, per end-user) for CVC Class 1 and 2 respectively on top of the existing AVC and CVC costs.  CVC Class 0 will have no additional charge.

The following table shows the AVC + CVC Class Fee. (Note: this is not representative of the cost of NBN Satellite plans.  Actual costs are likely to be higher as this excludes any CVC, NNI, network transit and peering costs incurred by Service Providers).

CVC Class CVC Class Fee AVC TC-4
12/1 Mbps
(12/1 Mbps)
25/5 Mbps
(25/5 Mbps)
0 $0.00 $24.00 $24.00 $27.00 $27.00
1 $18.00 $24.00 $42.00 $27.00 $45.00
2 $40.00 $24.00 $64.00 $27.00 $67.00

Data usage caps

On the data usage cap side, the company is two key enforcement metrics across all classes.  As part of the Fair Use Policy, nbn will introduce a per-AVC (essentially, per end-user) cap calculated on a rolling 4 week average basis.  The maximum usage over the four weeks are 75 GB, 100 GB and 150 GB for CVC Class 0, 1 and 2 respectively.  This is compared to 50 GB currently being enforced on the nbn Interim Satellite Service.

A second data usage enforcement metric is the “Peak Period” usage, which is calculated as an average across the Service Provider’s CVC Class user base.  During this peak period, defined as being between 7:00 am to 1:00 am (the following day) in the local timezone of the premises — the average upload/download data usage across all the ISP’s CVC Class users is 15/3GB, 20/4GB and 25/5GB for CVC Class 0, 1 and 2 respectively.

(25/5 Mbps)*
User data cap
(4-week rolling avg)
Peak Period Download
(ISP 4-wk average)
Peak Period Upload
(ISP 4-wk average)
0 $27.00 75 GB 15 GB 3 GB
1 $45.00 100 GB 20 GB 4 GB
2 $67.00 150 GB 25 GB 5 GB

* this is not representative of the cost of NBN Satellite plans.  Actual costs are likely to be higher as this excludes any CVC, NNI, network transit and peering costs incurred by Service Providers.

Failure for the service provider to comply with the data usage caps could result in data deprioritisation by nbn and/or shaping upload and download speeds to 256/256kbps until the issue is resolved.

Contention ratios

nbn will also place restrictions on the CVC-to-AVC contention ratio, enforcing a minimum number of end users a service provider must have on a particular CVC before allowing them to step up to the next tier.

On a CVC Class 0 CVC, the CVC capacity can range from the initial 100 Mbps up to 475 Mbps in 25 Mbps intervals.  The service provider must have at least 890 AVCs (equivalent to users) in that CVC Class in order to increase their CVC capacity beyond the initial 100 Mbps. A minimum of 210 additional AVCs are required for each 25 Mbps capacity increase.

For CVC Class 1, the maximum CVC capacity is 550 Mbps.  To move beyond the initial 100 Mbps CVC, the service provider must have at least 680 AVCs and can increase in 25 Mbps intervals for every 195 AVCs.

CVC Class 2 CVCs have a maximum CVC capacity is 725 Mbps.  Moving beyond the initial 100 Mbps CVC would require 548 AVCs and 25 Mbps increases for every 142 AVCs.

CVC Class Maximum CVC (Mbps) Minimum AVCs
(end-users) required
to move beyond
initial 100 Mbps
AVCs (end-users)
required for each
25 Mbps interval
beyond 100 Mbps
0 475 890 210
1 550 680 195
2 735 548 142

For assuming the 12/1 Mbps tier across all users: CVC class 0 provides a minimum contention ratio of ~1:106 with CVC class 1 and CVC Class 2 requiring a minimum contention of ~1:81 and ~1:68 respectively.

thoughts & analysis

Managing satellite network traffic has always been a tricky issue.  One one hand, the company doesn’t want to be perceived as disadvantaging rural and regional areas.  On the other hand, it is a necessity to make sure the service isn’t so heavily congested that no-one can use it.

The Interim Satellite Service was a good learning experience for nbn.  Service providers wanted to offer their users “unlimited” or high quota plans, thinking it could be offset by many more low-usage users.  They soon realised that the providers were unable to manage their own capacity well enough and soon, implemented traffic management policies to help manage the overly congested 48,000 users on the limited existing capacity of Optus and IPStar.

The limitations set by nbn are complex, there’s no doubt about that.  There are more intricacies that are omitted in this post — simply because they are too hard to explain simply.  These complexities means that small RSPs may struggle with implementing the traffic limiters and monitoring tools required to meet nbn‘s strict policies.  This could result in reduced competition in the long run.

On the quota…

Read updated analysis

Note from the writer: Please take this section with a grain of salt.  My views have changed substantially since this section was written because I had assumed that an end user could actually access all of the 75GB, 100GB and 150GB of data on CVC Class 0, 1 and 2 respectively.  However, it is now evident to me that the Fair Use Policy has implications on time of use and doesn’t account for data incurred during shaped data usage.  Please read the updated analysis for more information.

So, will 75 GB on a rolling 4-week average be enough? As a standard quota, I think it will fit within most people’s needs.

Currently, NBN offers 50 GB on a 4-week rolling average to their Interim Satellite Service.  So, 75 GB may seem like a small bump — but it’s 50% more nonetheless.

To keep things in perspective, statistics released by ABS show the average data consumption across all fixed-line technologies (end of 2014) is around 60 GB.  Further to that, the statistics released by nbn in their end of FY14/15 shows average data consumption on the NBN has steadily risen from 80 GB in September 2014 to 109 GB in June 2015.  So on an average consumption basis, I think the 75 GB monthly quota would fit the bill for most people.

So while nbn would expect most customers who take up a standard-tier 75 GB plan, the option is there to bump it up to 150 GB which is technically above the 109 GB average for current fixed-line/fixed-wireless nbn end users.

We also need to consider the demographic of the majority of these users: they live in very rural or remote communities.  Most regional centres are serviced by NBN Fixed Wireless and some form of NBN Fixed-line technology.  The remaining of these customers would have previously had very limited Internet connectivity, if any.

The standard 75 GB, plus the option of the more expensive 100 and 150 GB tiers will provide these rural communities with more than just a taste of online connectivity.  It will give users access to services like email and YouTube, and perhaps even more revolutionary — access to video conferencing through the likes of Skype or access to massive libraries of online TV content on Netflix, Presto or Stan.  All this will be possible within the confines of 75 GB, provided image quality is throttled accordingly.

Ideally, of course, each user should be given more capacity.  Hopefully, with the launch of the second satellite, nbn is able to augment some of the redundancy capacity on that satellite to use as more bandwidth for each user.

Relevant documents

[Source: nbn Testing Arrangements, Satellite Trial Product Terms supporting documentation]

Read More

NBN Long Term Satellite beams

NBN to cease Interim Satellite orders mid-November

Remote customers advised to wait until Long Term Satellite launch in the second quarter of 2016.

The company responsible for building the National Broadband Network, nbn, has updated their public website this afternoon — informing customers wishing to connect to their current Interim Satellite Solution (ISS) and the nbn™ Satellite Support Scheme (NSS) that it will cease accepting new orders for both services on November 15th, 2015.  nbn will also cease installation of new equipment from December 15th, 2015.

This comes as the first NBN satellite, Sky Muster, successfully launched at the start of this month and is expected to commence commercial services in the second quarter of 2016.

The company has advised would-be customers of the ISS and NSS that after the cease-sale dates for both services:

[…] you will need to wait for the launch of our Long Term Satellite services expected in the second quarter of 2016 and then connect your retail service provider.

The Interim Satellite Solution (ISS) was a temporary service set-up to replace the existing Australian Broadband Guarantee (ABG) while the new NBN satellites were being designed and launched.  The service, which uses existing satellite capacity on Optus and IPStar satellites, reached its maximum capacity of 48,000 customers in mid-December 2013 when it saw higher-than-anticipated take-up and use — leading to a serious congestion during peak periods.  The company had since implemented a number of changes, including enforced data caps, to prevent service abuse.  The service was opened up to an additional 9,000 activations in April 2014.

The new Long Term Satellite (LTS) significantly upgrades the existing satellite services on both the ISS and NSS, and is designed to deliver speeds of up to 25 Mbps download and up to 5 Mbps upload to customers in remote and regional Australia.

[nbn™ Satellite Support Scheme]

Read More


NBN sets December launch date for up to 50/20Mbps speeds

Product construct similar to NBN’s copper network

The company responsible for building the National Broadband Network, nbn, has released an updated Wholesale Broadband Agreement (WBA) to service providers which includes the new up to 50/20 Mbps speed range.

Originally, the company had limited the speed achievable on the NBN Fixed Wireless network to be 25/5 Mbps. However, according to the Product Technical Specification document, the equipment (W-NTD) installed in homes are capable of speeds of up to 75 Mbps downstream and 20 Mbps upstream depending on the equipment version.

nbn will begin to offer the 25-50/5-20 Mbps speed tier for the commonly used Traffic Class 4 Access Virtual Circuit, starting 1st December 2015. The company will also offer a 0.3 Mbps Traffic Class 1 Access Virtual Circuit — up from 0.15 Mbps, intended for low-latency services such as VoIP lines.

However, unlike the other Fixed-line technologies such as Fibre, FTTN and HFC — the NBN Fixed Wireless network will not offer an up-to 25/10 Mbps speed tier.

The company has been trialing the product since mid-May this year.

Read More

Graph showing planned premises in the 3 year rollout plan sorted by electorate margin, normalised by dividing the number of premises by the number of seats held

Analysis: cross-bench and marginal seats benefit from 3 year NBN plan

Rollout numbers are higher in the densely populated areas held by cross-bench and marginal LNP MPs.

Writer’s note, 20th Oct: A point of clarification. While the data does show on an averaged party-margin basis that seats of a few cross-bench MPs have a higher number of homes in the 3 year plan, this does not suggest a direct correlation between the two. As stated previously in the article, albeit possibly not clearly enough, the seats held by the Greens, KAP and PUP are generally more densely populated and thus, would tend to have higher rollout figures in the fixed-line 3 year plan.

Analysis of the nbn construction plan released last week shows that on an averaged per-electorate basis, seats held by cross-bench MPs Adam Bandt (Greens), Bob Katter (Katter Australian Party), Clive Palmer (Palmer United Party) and conservative LNP seats has the greatest number of premises scheduled to commence construction over the next 3 years. However, the majority of these seats are coincidentally located in the more densely populated areas of Australia.

By party margin on a per-electorate basis

Between now and September 2018, Adam Bandt’s electorate of Melbourne is expected to see 99,600 premises commence build.  Bob Katter’s electorate of Kennedy is expected to see 71,500 premises commence build, followed by Clive Palmer’s electorate of Fairfax taking 68,200 premises into build by late 2018.

On a per-electorate basis (that is, the total number of premises divided by the number of seats held), marginal Liberal-National Party and Nationals Party seats come in at fourth and fifth with 59,233 and 56,300 premises expected to enter build respectively.

Party Margin classification Total planned fixed-line Electorates under this category Planned Fixed-line (per-electorate)*
GRN Cross bench  99,600 1  99,600
KAP Cross bench  71,500 1  71,500
PUP Cross bench  68,200 1  68,200
LNP Marginal  355,400 6  59,233
NAT Marginal  56,300 1  56,300
ALP Safe  836,200 15  55,747
LIB Fairly safe  493,500 9  54,833
LIB Safe  1,430,800 29  49,338
ALP Marginal  1,413,400 29  48,738
NAT Fairly safe  48,600 1  48,600
NAT Safe  331,500 7  47,357
LNP Safe  323,200 7  46,171
LNP Fairly safe  406,200 9  45,133
ALP Fairly safe  472,500 11  42,955
LIB Marginal  819,500 20  40,975
IND Cross bench  60,500 2  30,250

* the total number of premises in the plan to commence build (column 3), divided by the number of seats held within the party-margin category (column 4)

Graph showing planned premises in the 3 year rollout plan sorted by electorate margin, normalised by dividing the number of premises by the number of seats held
Graph showing planned premises in the 3 year rollout plan sorted by seat margin, normalised by dividing the number of premises by the number of seats held in each category

By electorate

However, the winning electorate is the electorate of Sydney – currently held by Labor MP Tanya Plibersek – with a total of 106,800 premises slated to commence build by late 2018.  This is followed closely by Adam Bandt’s seat of Melbourne (99,600), Labor MP Melissa Parke’s marginal seat of Fremantle (94,700) and Labor MP Clare O’Neil’s seat of Hothan (92,000).  The high number of premises in these areas is reflective of the dense population of these inner metropolitan centres.

Of the 150 federal electorates, only the electorate of Solomon – held by Country Liberal MP Natasha Griggs – has no fixed-line rollout plans.  The electorate covers the greater Darwin area where NBN rollout is mostly completed or already underway.

Electorate Name MP elected in 2013 Held by Margin classification Planned FTTx Premises Planned HFC premises Planned Fixed Line (total)
Sydney Tanya Plibersek ALP Safe 12,100 94,700 106,800
Melbourne Adam Bandt GRN Cross bench 24,800 74,800 99,600
Fremantle Melissa Parke ALP Marginal 89,500 5,200 94,700
Hotham Clare O’Neil ALP Fairly safe 6,000 86,000 92,000
Flinders Greg Hunt LIB Safe 90,900 0 90,900
Higgins Kelly O’Dwyer LIB Fairly safe 0 90,000 90,000
Bruce Alan Griffin ALP Marginal 29,500 54,600 84,100
Wakefield Nick Champion ALP Marginal 55,100 28,900 84,000
Leichhardt Warren Entsch LNP Marginal 63,500 19,600 83,100
Chisholm Anna Burke ALP Marginal 700 80,600 81,300
Moncrieff Steven Ciobo LNP Safe 17,600 63,600 81,200
Warringah Tony Abbott LIB Safe 18,000 60,000 78,000
Kooyong Josh Frydenberg LIB Safe 0 75,600 75,600
Batman David Feeney ALP Safe 1,700 73,200 74,900
Blaxland Jason Clare ALP Safe 4,300 69,900 74,200
Kingsford Smith Matt Thistlethwaite ALP Marginal 1,600 72,400 74,000
Corio Richard Marles ALP Fairly safe 72,800 0 72,800
Kennedy Bob Katter KAP Cross bench 59,100 12,400 71,500
Fairfax Clive Palmer PUP Cross bench 68,200 0 68,200
Stirling Michael Keenan LIB Safe 59,800 7,800 67,600
Deakin Michael Sukkar LIB Marginal 2,500 63,800 66,300
Cowan Luke Simpkins LIB Fairly safe 43,000 22,200 65,200
Adelaide Kate Ellis ALP Marginal 17,700 47,000 64,700
Jagajaga Jenny Macklin ALP Marginal 8,500 55,500 64,000
Forrest Nola Marino LIB Safe 63,500 0 63,500
Corangamite Sarah Henderson LIB Marginal 63,400 0 63,400
North Sydney Joe Hockey LIB Safe 7,900 55,500 63,400
Boothby Andrew Southcott LIB Fairly safe 25,600 37,000 62,600
Bendigo Lisa Chesters ALP Marginal 61,900 0 61,900
Lalor Joanne Ryan ALP Safe 32,500 29,300 61,800
Fraser Andrew Leigh ALP Safe 48,400 13,400 61,800
McEwen Rob Mitchell ALP Marginal 61,500 0 61,500
Bowman Andrew Laming LNP Fairly safe 36,100 25,000 61,100
Dunkley Bruce Billson LIB Marginal 36,100 24,100 60,200
Port Adelaide Mark Butler ALP Safe 51,600 8,200 59,800
Greenway Michelle Rowland ALP Marginal 23,300 36,000 59,300
Bradfield Paul Fletcher LIB Safe 0 58,900 58,900
Petrie Luke Howarth LNP Marginal 17,200 41,600 58,800
Gellibrand Tim Watts ALP Safe 14,100 44,700 58,800
Gorton Brendan O’Connor ALP Safe 24,900 33,400 58,300
Chifley Ed Husic ALP Safe 9,000 49,000 58,000
Isaacs Mark Dreyfus ALP Marginal 21,300 36,400 57,700
Fisher Mal Brough LNP Fairly safe 57,600 0 57,600
La Trobe Jason Wood LIB Marginal 43,500 13,900 57,400
Mitchell Alex Hawke LIB Safe 26,800 30,000 56,800
Gippsland Darren Chester NAT Safe 56,400 0 56,400
Page Kevin Hogan NAT Marginal 56,300 0 56,300
Richmond Justine Elliot ALP Marginal 55,900 0 55,900
Canning Don Randall LIB Safe 55,800 0 55,800
Farrer Sussan Ley LIB Safe 55,700 0 55,700
Gilmore Ann Sudmalis LIB Marginal 55,200 0 55,200
Tangney Dennis Jensen LIB Safe 44,100 11,100 55,200
Paterson Bob Baldwin LIB Fairly safe 55,100 0 55,100
Bonner Ross Vasta LNP Marginal 8,400 46,600 55,000
Capricornia Michelle Landry LNP Marginal 54,800 0 54,800
Forde Bert van Manen LNP Marginal 28,800 25,800 54,600
Barton Nickolas Varvaris LIB Marginal 1,700 52,800 54,500
Cook Scott Morrison LIB Safe 5,900 48,200 54,100
Dickson Peter Dutton LNP Fairly safe 20,200 33,500 53,700
New England Barnaby Joyce NAT Safe 39,800 13,900 53,700
Calare John Cobb NAT Safe 53,500 0 53,500
Moreton Graham Perrett ALP Marginal 5,000 47,600 52,600
Throsby Stephen Jones ALP Fairly safe 52,500 0 52,500
O’Connor Rick Wilson LIB Marginal 52,300 0 52,300
Barker Tony Pasin LIB Safe 51,700 0 51,700
Riverina Michael McCormack NAT Safe 51,600 0 51,600
Sturt Christopher Pyne LIB Safe 18,800 32,200 51,000
Hunter Joel Fitzgibbon ALP Marginal 50,800 0 50,800
Hume Angus Taylor LIB Safe 49,000 1,600 50,600
Mayo Jamie Briggs LIB Safe 45,900 4,700 50,600
Dawson George Christensen LNP Fairly safe 50,300 0 50,300
Kingston Amanda Rishworth ALP Fairly safe 47,100 3,100 50,200
Indi Cathy McGowan IND Cross bench 50,100 0 50,100
Wills Kelvin Thomson ALP Safe 6,100 43,700 49,800
Durack Melissa Price LIB Marginal 49,500 0 49,500
Brisbane Teresa Gambaro LNP Marginal 4,000 45,100 49,100
Fadden Stuart Robert LNP Safe 23,100 26,000 49,100
Grey Rowan Ramsey LIB Safe 48,900 0 48,900
Reid Craig Laundy LIB Marginal 12,800 36,000 48,800
Mallee Andrew Broad NAT Fairly safe 48,600 0 48,600
Pearce Christian Porter LIB Fairly safe 43,000 5,400 48,400
Bennelong John Alexander LIB Fairly safe 2,600 44,800 47,400
Aston Alan Tudge LIB Fairly safe 2,200 44,800 47,000
Ryan Jane Prentice LNP Fairly safe 13,000 34,000 47,000
McMillan Russell Broadbent LIB Safe 46,900 0 46,900
Grayndler Anthony Albanese ALP Safe 1,200 45,400 46,600
Eden-Monaro Peter Hendy LIB Marginal 46,500 0 46,500
Flynn Ken O’Dowd LNP Fairly safe 45,300 0 45,300
Wannon Dan Tehan LIB Safe 45,300 0 45,300
Perth Alannah MacTiernan ALP Marginal 45,200 0 45,200
Oxley Bernie Ripoll ALP Marginal 23,800 21,300 45,100
Shortland Jill Hall ALP Fairly safe 44,900 0 44,900
Lyne David Gillespie NAT Safe 44,600 0 44,600
Brand Gary Gray ALP Marginal 44,600 0 44,600
Makin Tony Zappia ALP Marginal 28,600 15,900 44,500
Calwell Maria Vamvakinou ALP Safe 29,800 14,300 44,100
Hinkler Keith Pitt LNP Fairly safe 43,500 0 43,500
McPherson Karen Andrews LNP Safe 12,600 30,600 43,200
Parkes Mark Coulton NAT Safe 43,200 0 43,200
Newcastle Sharon Claydon ALP Fairly safe 43,200 0 43,200
Maranoa Bruce Scott LNP Safe 42,100 0 42,100
Casey Tony Smith LIB Fairly safe 25,400 16,600 42,000
Braddon Brett Whiteley LIB Marginal 39,900 0 39,900
Wentworth Malcolm Turnbull LIB Safe 0 39,600 39,600
Banks David Coleman LIB Marginal 1,100 38,300 39,400
Curtin Julie Bishop LIB Safe 1,500 37,100 38,600
Holt Anthony Byrne ALP Fairly safe 19,200 19,300 38,500
Rankin Jim Chalmers ALP Marginal 4,700 33,100 37,800
McMahon Chris Bowen ALP Marginal 8,000 29,200 37,200
Wide Bay Warren Truss LNP Safe 36,500 0 36,500
Groom Ian Macfarlane LNP Safe 35,900 0 35,900
Swan Steve Irons LIB Fairly safe 33,300 2,500 35,800
Canberra Gai Brodtmann ALP Fairly safe 35,400 0 35,400
Macarthur Russell Matheson LIB Safe 35,200 0 35,200
Wright Scott Buchholz LNP Safe 35,200 0 35,200
Hasluck Ken Wyatt LIB Marginal 34,700 0 34,700
Murray Sharman Stone LIB Safe 34,700 0 34,700
Moore Ian Goodenough LIB Safe 9,900 24,600 34,500
Goldstein Andrew Robb LIB Safe 0 34,300 34,300
Blair Shayne Neumann ALP Marginal 14,400 18,400 32,800
Hindmarsh Matt Williams LIB Marginal 6,100 24,100 30,200
Griffith Kevin Rudd ALP Marginal 7,400 22,800 30,200
Maribyrnong Bill Shorten ALP Safe 1,600 28,200 29,800
Lilley Wayne Swan ALP Marginal 3,600 25,800 29,400
Cowper Luke Hartsuyker NAT Safe 28,500 0 28,500
Robertson Lucy Wicks LIB Marginal 28,400 0 28,400
Mackellar Bronwyn Bishop LIB Safe 12,100 15,800 27,900
Melbourne Ports Michael Danby ALP Marginal 3,700 24,000 27,700
Scullin Andrew Giles ALP Safe 12,900 14,800 27,700
Macquarie Louise Markus LIB Marginal 27,000 0 27,000
Lyons Eric Hutchinson LIB Marginal 26,900 0 26,900
Lindsay Fiona Scott LIB Marginal 5,100 21,500 26,600
Ballarat Catherine King ALP Marginal 25,200 0 25,200
Hughes Craig Kelly LIB Safe 5,000 19,900 24,900
Herbert Ewen Jones LNP Fairly safe 24,600 0 24,600
Werriwa Laurie Ferguson ALP Marginal 8,300 15,900 24,200
Berowra Philip Ruddock LIB Safe 7,200 16,600 23,800
Fowler Chris Hayes ALP Safe 9,800 14,000 23,800
Longman Wyatt Roy LNP Fairly safe 23,100 0 23,100
Cunningham Sharon Bird ALP Fairly safe 20,900 0 20,900
Parramatta Julie Owens ALP Marginal 3,600 13,300 16,900
Menzies Kevin Andrews LIB Safe 0 16,800 16,800
Franklin Julie Collins ALP Marginal 14,800 0 14,800
Charlton Pat Conroy ALP Fairly safe 12,900 0 12,900
Lingiari Warren Snowdon ALP Marginal 11,300 0 11,300
Denison Andrew Wilkie IND Cross bench 10,400 0 10,400
Watson Tony Burke ALP Fairly safe 3,500 5,700 9,200
Dobell Karen McNamara LIB Marginal 8,300 0 8,300
Bass Andrew Nikolic LIB Marginal 4,000 0 4,000
Solomon Natasha Griggs CLP Marginal 0 0 0

Read More