Opinion: NBN and the art of planning

 

Expect NBN Co's plan to be wrong... for the right reasons.

The NBN Business plan, like Australia’s response to floods in Queensland, provides insights into the art of planning writes David Havyatt. 

We Australians are a pretty cynical lot.  Despite having above average health and education outcomes in the OECD for below average expenditure, we tend to bemoan the state of our hospitals and schools. 

We tend to raise concerns about inflationary pressure, for example, in response to Government spending programs proven to have boosted consumer confidence during the global financial crisis.

Whatever your take on these matters - the difference between a government bumbling through such challenges and achieving successful outcomes is usually down to good planning. 

Just before Christmas Australians were afforded the opportunity to read NBN Co’s 2011-2013 Corporate Plan.  The release of this plan had been delayed due to the need to amend it to reflect the determination of the ACCC of the number of Points of Interconnection required for a competitive market.

In my opinion it is a very robust plan, very soundly based on a methodology that any corporation would recognise. 

If you’ll allow me to digress, a slightly different kind of planning has been in evidence in response to the floods in Queensland.  Overseas news reports like the New York Times have really put these floods into context by describing the area flooded as about the size of France and Germany. 

Anna Bligh was very composed discussing the State Government response to this disaster. In particular she mentioned how changes had been made to command structures following Cyclone Larry.  She also mentioned the success of the SMS alert system, a national initiative following last year’s bushfires.

Similarly Australia’s Prime Minister noted at a press conference that offers of assistance have come from New Zealand and the USA, but that we don’t actually need any help right now.  

That’s because flooding is something we expect, and no amount of attempts at damming could ever stop it. Australian Governments at all levels plan for events like these.  And if the response in Queensland is any guide, ultimately they do it well.

Premier Bligh noted that sometimes a post-event review suggests ways to do things better, so you adapt and prepare again.

That brings us back to the NBN Co Corporate Plan.  The plan covers just three years of a 9.5 year deployment and thirty year investment.  From there on the plan does all the things you would expect; it identifies the objectives and the existing environmental constraints, it describes the current and future market and outlines the steps to be taken to start achieving the objectives. 

The plan recognises that the future cannot be perfectly predicted and hence deals with certain risks and uncertainties.  Some of the planning assumptions that constitute risks are significant for overall financial viability, but the majority of these (like policy settings on cherry-picking and planning) are within the control of NBN Co’s shareholders.

The area of risk that is outside NBN Co and government control is market size.  Here NBN Co has actually been quite conservative.  NBN Co chief executive Mike Quigley used this term to describe described the assumption that only 70 percent of homes passed will take up the fixed service. 

The plan included a graphic of download speeds since 1985.  This really makes the case that the NBN isn’t needed now … but it is needed at the point of its completion in ten years time. 

 

 

I find it interesting that those who claim we don’t need to start building the network now also question the ability to achieve the forecast number of connections.  The peak of over one million connections in a year equates to 20,000 a week or 4,000 every working day.  But if we don’t start now the connection task becomes greater.

 

The plan is also robust in undertaking a review of available technologies, not just making the presumption  that the FTTH goal of Government should be beyond analysis.

The plan has had to accommodate the Government’s objective of uniform national wholesale prices.  The real goal is nationally uniform retail prices - which others have argued could be achieved simply through market competition.

For example, prior to competition in long distance calling, charging was based on a series of distance based charge bands.  Critics of competition complained it would result in cheaper calls on thick routes and dearer calls on thin ones. 

The outcome was completely different - it has been the elimination of price discrimination on any long distance call.  So a call from Mt Isa to Kalgoorlie costs as much as a call from Sydney to Melbourne despite being made up of PSTN originating and terminating access charges up to four times higher at each end, and having three long transmission legs (Mt Isa to Brisbane, Brisbane to Perth, Perth to Kalgoorlie), compared with one.  Two of these are over long, thinly-used routes.

Analysis of cases like this and the existing fixed broadband market can show policy makers how to use markets to get outcomes.  Thankfully the ACCC saw fit to insist on more points of interconnect.

The most interesting part of commentary on the business plan has been about the rate of return, 7.04 percent.   It is interesting because it is in part misguided.  The low rate of return isn’t so much an outcome of the plan, as an input to it.  Item four of the list of objectives “enhanced by correspondence from the Government” is that the expected rate of return should, at a minimum, be in excess of the current public debt rates.

NBN Co has planned to the constraint.  To be getting a higher return than requested by the funder of the plan would be to do the one thing that a return to a monopoly provider would risk, that is excessive returns.  Over the long run as citizens we don’t want NBN Co to return to the Government anything more than it has cost to build the network. That wasn’t the point of the exercise.

The NBN Co plan is robust and provides a convincing guide to how the NBN will be delivered.  The one thing we can be sure of is that the plan will be wrong – there will have to be some things that won’t turn out the way it has been represented.  It isn’t out of the question that demand and need for NBN services far exceeds the plan’s expectations and there is a need to accelerate deployment.

The challenge here will be to show the same kind of skill as has been displayed in Australia’s disaster planning and to adjust the plan in the face of real outcomes. 

Copyright © iTnews.com.au . All rights reserved.


Opinion: NBN and the art of planning
David Havyatt
"@Umbria…. Hmmm yes. From a pure technical perspective you are correct that multicast reduces bandwidth requirements. However when the commercial dynamics are overlaid, the existence of ..."
By adavion
 
 
 
Comments: 31
sydneyla
Jan 7, 2011 7:46 AM
David you really do sound like a pro Labor dreamer. As a Queenslander I do not see your projected marvellous Labor solution for our flood disaster. Why did you not mention the diabolical foul-up of the burning bats horror or the wasteful school building scheme?

Everybody knows that a NBN would be a good idea so would a super highway around Australia, but the question being asked is if fast broadband could be delivered at a cheaper cost? Also, with the extra cost involved how many Australians will want the maximum speed offered?
hellfire
Jan 7, 2011 8:18 AM
I agree with sydneyla and question why the NBN has to be a government project when anything the government carries out is at a far greater cost to the same outcome done by the private sector. The present government has not prooved to be a good money manager of any of it's programs so far so why would the NBN be any different?
I also question the need for the NBN at all when only a very small percentage of Australians do not have access to an adequate broadband connection and this could easily be delivered at far less cost by wireless technology which is fast becoming the way of the future as more and more users like the flexability of a connection that you can take with you anywhere.
Rossyduck
Jan 7, 2011 8:51 AM
At a more basic level and referring to the Qld floods as well. With this NBN co one size fits all design - intrinsically appealing to those who do not know what they are talking about but like to nod their heads sagely at the thought of a uniform network across the country, the entire Qld area under flood now would have had to be redone !. Every FDH serving 200 homes would have to be replaced, every connector in every FDH would need to be stripped out and respliced. Under the NBN Co one size fits all the entire region would have been without comms (copper retired) for months. I use this as one example. Far better to have had smaller businesses who know an area and build to those peculiarities.
djzort
Jan 7, 2011 9:19 AM
A super highway around Australia would be a fantastic alternative to NBN. Anyone drive to Melbourne from Sydney lately? ...
deonast
Jan 7, 2011 9:31 AM
@Rossyduck hold on so you are saying that fibre infrastructure is not resistant to food damage, do you have any technical knowledge to back this up. Currently we don't see telstra replacing the whole POTS network everytime there is a flood.
Ice
Jan 7, 2011 9:58 AM
@ sydneyla the burning bats fiasco as you call it is a media beet up dozens of home were burning down across each state every year due to poorly installed batts way before any scheme by the government.

I am qualified in the building industry, and have looked in ceilings in horror at the way batts have been installed for at least 15 years, advising people to get them reinstalled but 99% of people in australia will not listen because they think they know better because everything in the media is true, instead of finding the facts out from people with experience.
cmasiero
Jan 7, 2011 10:22 AM
Hi _is_ a pro labor dreamer. Have you seen his blog, it's a veritable fan page. Try some critical thinking there David!
serrendipity
Jan 7, 2011 10:26 AM
I agree with the earlier respondents - the issues with NBN are in relation to the technology to be used and the cost. One size fits all is very expensive and how many Australians want or need the maximum speed? There is no debate on the fact that national broadband is a good idea. However, as with all IT projects, the need should have been better assessed and the solution provided to fit the need in a cost effective manner for the taxpayers. The current plan can be compared with providing an IBM mainframe computer system (to prepare for the future) for a corner store at a cost in the millions of dollars when a standard $599 PC would provide the capacity the store needs for the next 5+ years. Govt still needs to reassess and justify the
NBN proposed technology plan before they commit to this elaborate expenditure.
mad1k5
Jan 7, 2011 10:57 AM
I disagree with you lot, simply because you replace one technology fixed line, with another fixed line technology.

IN the technology Industry you cannot apply the same RULES to a expensive computer to a cheap computer, or video cards and etc.

I think alot of people have lost the plot.

Alot of the core cables such as backhaul, last mile actually stay in the ground longer than most of the other components of the network.

@serrendipity, then disconnect your copper line then.
Graeme Harrison (prof at-symbol post.harvard.edu)
Jan 7, 2011 11:28 AM
As I've posted here before, the BIG NBN issue not dealt with before (or above) is the sensitivity analysis. As a former senior management consultant I can say that McKinsey should never have put their name to any analysis that did not do a proper sensitivity analysis. The big issue is where you cut-over from fibre to wireless, and if you bring the fibre coverage down from 93% to 90% then 85%, the thing gets so positive in returns, no-one would have an issue. The question will always be whether long runs to outer-urban 10-acre blocks makes sense, if those customers have indicated no desire to be hooked-up. The business centres and major population hubs are a no-brainer, but in time we will realise that we should delay fibre to marginal consumption areas.. and if that means a better return and a 20-year roll-out, so be it.

I'm all in favour of uniform pricing nationally, but believe that the technology breakeven decisions for very marginal areas are being ideologically driven, rather than thought through with economics in mind. Every run of useless fibre (few or no customers on long-runs between properties) is another hospital bed we can't have, or renewable energy project we need to defer!
Tailgator
Jan 7, 2011 12:20 PM
@ 'Prof Graeme Harrison - Harvard'
- search reveals no such staff member https://www.directory.harvard.edu/phonebook/submitSearch.do;jsessionid=GC6dNmpcHgsTPvhLnmYtmn3n1pNd1qcSCVh8y5TLSrB2z21Vc26y!969321850
- another hospital bed 'we' can't have, or renewable energy project 'we' need to defer! Who's we?

Pull the other one.
Tailgator
Jan 7, 2011 12:24 PM
What a lot of waffle from sydneyla onwards, dragging up the same old misinformed, biased, and factually incorrect rhetoric. Bats, Ber, private sector does it cheaper, don't need the speed, wireless is better, etc, etc. On and on it goes as if the simple act of repetition will make it true.

Yet again, ....

1. Bats, Ber, the old touchstones for the pro Liberal tribe. Yawn. Read the inquiry reports.
2. Private? Faster broadband to a standard, degree of ubiquity, and future proofing could not be delivered by anything other than the NBN.
3. Wireless will not be able to meet the demands in both the intermediate and long term future. Every industry technologist of note supports this view. And even now demand for data on fixed line is increasing at a far greater rate than that of wireless (which is decreasing) despite the higher rate of new connections. (Check the Bureau of Stats.
4. Demand for data is increasing exponentially. The capacity may not be fully utilitized for another 10 yrs (about the time NBN will done) but at least we'll be ready for it.
5. The private sector does it cheaper ??? What is it about required rates of return that is misunderstood. If in doubt check Telstra's proposed pricing for it's 2006 FTTN plan in the cherry picked metropolitan areas. The NBN will deliver better access at prices no private sector would or could.
6. Floods/local knowledge, so the copper network will 'just work' once the water subsides ??
7. @ Prof from Harvard. (lol)
- How much is enough? Who decides?
- The whole purpose is not to make a massive profit. It is about providing infrastructure of the best possible type, for as far into the future as possible and at the least cost to the citizens (who happen to be the consumers) in this country.

There are none so blind as those who will not see, and none so blind as those blinded by their own political/ideological faith.
HubertCumberdale
Jan 7, 2011 12:44 PM
hellfire wrote:
I also question the need for the NBN at all when only a very small percentage of Australians do not have access to an adequate broadband connection

Do you consider less than 1mbit upload to be an "adequate broadband connection"?

hellfire wrote:
this could easily be delivered at far less cost by wireless technology which is fast becoming the way of the future as more and more users like the flexability of a connection that you can take with you anywhere.

No.
peterh_oz
Jan 7, 2011 2:38 PM
Ok time for me to weigh in ...

"The current plan can be compared with providing an IBM mainframe computer system (to prepare for the future) for a corner store at a cost in the millions of dollars when a standard $599 PC would provide the capacity the store needs for the next 5+ years."

You don't build a network for 5 years. You build it for 50 years. And there is no point waiting for the $599 PC to be full, too slow and unusable before you start with a replacement - especially if the replacement takes 10 years!

That is the problem with this Country - everyone is a critic (sometimes justified), and the pollies are looking at 3-4 year windows to win the next election, not building or planning long term for the country.

FINALLY someone had the guts and "courage" to say enough is enough, the copper which is already failing many people will be near useless to far too many people (yes, not everyone!) in 10-15yrs, so lets start a long term plan to replace it.

There are 2 ways - replace the bits in the most expensive, under serviced rural areas, and keep the HFC and ADSL2 in the SMALL portion of the cities where it is at or near top speed (12M or more), OR replace the whole lot and use the income generated from the cities to cross-subsidise the bush thus resulting in lower costs overall.

Oh and a $24 wholesale price as a starting point IS a lower cost for almost everyone, and I don't believe the $60 retail costs. Only the Tier 1 providers (Optus/Telstra) are charging more than double the wholesale price for entry level plans.

And remember, the phone only ie non internet plans will remain at current line rental prices, no more.

I voted for Howard and his GST, cos I knew we needed it regardless of the "what about me" incidents of some price rises, and I voted for the NBN - twice! I just hope it gets finished before the far-right wreckers (as Fraser says, they're not Liberal) wreck it.
umbria
Jan 7, 2011 3:45 PM
Happy New Year to the Muppet Show hecklers, SydneyLaLa, RossyDuck and Harvard Professor (actually a wireless tech salesman).

Rossyduck, the Qld floods would not have destroyed every NTU in an area the size of Germany and France. Only 300 houses suffered water ingress, so 300 NTUs would need checking and cleaning or replacement. The submeregd fibre would also remain completely functional, unlike the ADSL over copper which suffers increased signal-to-noise when the ducts fill with water.

Many also forget that fibre is only being used in dense large towns. If you bother to read all the NBNCo and Senate Committee documentation in the public domain for up to a year you will see that the mix of technologies in this proportion is also the cheapest way to deliver ubiquitous broadband. The NBN delivers minimum 12 Mbps guaranteed broadband for the first time to 40% of Australians (7 million of us) who have no option but dialup or expensive satellite after fifteen years of competition. For the 93% with fibre it delivers speeds selectable up to 1 Gbps.

It is indeed a no-brainer to build it.
FLashy
Jan 7, 2011 4:14 PM
Apart from the Government standard spin from this cost analysis document think of the infrastructure needed to maintain this new system of comms in a household.
The new modem and it's new power supply will generate another industry, let alone all the new PC's required to process the higher line speeds. The storage requirements will be terabytes and terabytes.
Think of the industry needed to supply trenching machinery and cherry pickers for overhead fibre cables.
Then we have the copper recovery which is another byproduct to be reused in the world.
The extra maintenance staff needed to fix all the breaks in the new fibre network will be substantial, currently we have minimal numbers as the current network is relatively sparse.
With 99.9% of Telstra's fibre underground, the cuts in services happen everyday but are under control, with overhead fibre cabling, the breakage will increase I believe because of more exposure.
Then there is the solar aspect of exposed fibre cables, will the jackets degrade under Australian UV-B conditions.
Nice to see that all the comments are about self interests and the usual political differences.
sydneyla
Jan 7, 2011 4:15 PM
umbria Seasons best wishes to you. I do understand your intense desire to have a fast broadband service but do you not think that your desire is somewhat selfish when the sick need hospitals, the poor need extra help financially and every other section of Australia need finance for their individual needs.

Sure, 100Mbps for all would be nice but so would an Aston Martin car be nice for every Australian but in reality perhaps a Holden would do. You are very tunnel vision focused umbria but please consider other Australians before you demand your personal goodies.
FLashy
Jan 7, 2011 4:32 PM
a quick note about an umbria comment.
"The submeregd fibre would also remain completely functional,"

The local fibre runs will have passive splitters which if underwater will increase the dB loss.
Can someone verify if the ONT's used have 'Ruggedised optical splitters"

The current air-gap is a 3dB loss I wonder what it will be with muddy water.
adavion
Jan 7, 2011 4:43 PM
@ Sydneyla - good question re cost - apparently Campbell Newman is doing it better (point to point) and cheaper (sewer-net with QUU)

@ Hellfire - In an ideal world a separated Telstra would implement the NBN. But do remember Telstra's cost of capital is around double the NBN's IRR - so it will still need to be government funded - Funding at the risk free rate is arguably dodgy but does allow the government to get its money back. Other than that, NBN Co is being run as a private company for as must profit as it can extract- subject (just like Telstra) to USO obligations. Bottom line NBN Co being 100% government owned makes no difference to costs and margins especially considering the likely 85% EBITDA margin -> most of the cost is in contracting international telecom vendors to build it.

@ Rossyduck - Good argument re FDH hidden in a completely illinformed example: NBN Co says it will offer 11,160Tb/s of fibre (1Gbps x 12m x 93%) but reading the fine print it is designed to deliver a 50kps CIR- slower than dial up. Data volumes are growing at >20%pa. McKinsey/KPMG says point to point will cost $4.5bn extra. It is just plain dumb to sacrifice 99.5% of capacity for around 10% of the CapEx.

@DJZort - provided its an Autobahn I'm in.

@Serrendipity - I need you (and Tony Abbott and all other NBN sceptics) to have the maximum speed in order to get off the net when I want to use it. i.e. put the contention bottlenecks as far back in the network as possible where they are least likely to shaft me. Also run a 25-35%pa growth rate in data volume out for 20 years or so and see how fast you will need to go when the NBN is built. Further, how do you explain the log-linear growth in speed from 1.2kps baud c1980 to 100mbps DOCSIS 3.0 today? Do you think this trend will immediately reverse itself?

Also the NBN is much more sensitive to the Cost of Capital than absolute CapEx. Allowing a 7% IRR and doubling the CapEx will allow for much cheaper broadband than halving the CapEx but doubling the cost of capital.


@Graham Harrison - I did like McKinsey / KPMG's sensitive analysis - 500 page document 4 pages of sensitivity generally +/- a couple of billion dollars or so. But they provided their CapEx estimate to $100m.

But what's wrong with ideologically driven investment- eg Schools, Hospitals, Police, Public Transport all fail to generate financial returns. note also if the total NBN investment > bond yield then the project pays for itself irrespective of how much money is wasted- so you don't actually miss out on hospital beds.

@Tailgator - Lol....

@Hubert - Agree.

@Peterh - The issue for me is not whether there is a subsidy, but whether the subsidy is designed in a manner that will impeed service improvement. McKinsey KPMG published the cost of transit backhaul vs CAN - with Transit BH costing around $5.4bn- around 13% of the build cost. Further they indicate the cost of upgrading to Point to Point is around $4.3bn.

However NBN Co expects to make over 33% of its income from transit backhaul- CVC pricing and the cost of upgrading from 12/1 to 1000/400 is 625%. This shows NBN Co proposes a massive cross subsidy from one part of the net to another. In a nutshell the Corporate Plan shows a company committed to delivering 50kps. Clearly NBN Co pricing is based on making 12/1 as cheap as possible, and with 240:1 contention 12Mbs-1 PIR is 50kps CIR.

[This sheds further light on its 14 POI interconnect plan (which would need 14, 800Tbs-1 transit backhaul cables to serve with 1:1 contention) as well, and its intent no to provide Layer 1 access].

In the world of regulated utilities the price charged must equal to the regulated return of the assets. Here NBN Co plans to breach that rule- thus consumers will be ripped off in the form of slower services.

By burying the subsidy within pricing, there is no independent benchmark on NBN efficiency and no incentive to provide a network that is optimised for performance.

While the NBN may be engineered to deliver 1Gbps PIR, the design of the CVC component with its implied 240:1 contention ratio implies the NBN will deliver a CIR of 4.17Mbs-1 to the point of interconnect. At a cost of $233.40 per month wholesale [$150 + 4.17x $20]. That is, 10x the cost of Telstra's copper for a speed slower than Telstra's average DSLAM connection.

@sydneyla - IRR of NBN > bond yield = self-funded project. Its the corporate finance definition of having buying your cake and having the original finance too. it too. This has been pointed out before- don't understand why you don't get it yet. Unless your argument is that take-up is overstated (seriously how many people will opt out when Telstra cuts their cables?) or CapEx is under stated (releasing the tender documents and ensuring overruns lie with the contractors will fix that).

I am also wondering if you have noticed how BMW and MB come out with new toys for their 7-Series / S-Class cars eg electric windows and ABS brakes that then filter down to plebmobiles like Holdens and Fords 10 years or so later? Go back and compare a 1960 Holden with the current model. It will take 10 years to build the NBN and it will have a useful life of at least 40 more. Then consider whether today's Aston Martin DBS, DB9 or Vanquish will be anywhere near good enough to be a drive car in 50 years time.
deteego
Jan 7, 2011 4:44 PM
Tailgator wrote:
@ 'Prof Graeme Harrison - Harvard'
- search reveals no such staff member https://www.directory.harvard.edu/phonebook/submitSearch.do;jsessionid=GC6dNmpcHgsTPvhLnmYtmn3n1pNd1qcSCVh8y5TLSrB2z21Vc26y!969321850
- another hospital bed 'we' can't have, or renewable energy project 'we' need to defer! Who's we?

Pull the other one.


http://www.health.gov.au/internet/nhhrc/publishing.nsf/Content/394/$FILE/394%20-%20SUBMISSION%20-%20Graeme%20Harrison.pdf

hmmm......
FLashy
Jan 7, 2011 4:51 PM
Another thought came to mind, has anyone made a study of the power, water and gas utilities that have been compromised in the Tasmanian optic fibre roll out?
Normally when contractors do trenching there is always damages to other services.
Dialling 1100 can not provide a 100% surety that 100 year old sewage and water pipes will not be affected.
In the older parts of our capital cities records are a bit thin on the ground and breaks will increase public anger.
But I suppose Mr Conroy will say it is for the overall good of the NBN.
Ace
Jan 7, 2011 11:36 PM
I notice a few people claiming that 'money used to build NBN could be used for hospital beds'. No. They are two entirely separate things. If money is not borrowed to spend on the NBN, then the money it not borrowed. There is no $35 billion sitting in a bank account waiting for the govt to decide how to spend it. Obviously.
Rossyduck
Jan 8, 2011 7:47 AM
Guys (including you Umbria),

Please read the NBN Co documentation before commenting.

The FHD is not the NTU - it is similar to the posts you see in the street distributing the copper. Yes - a copper cabinet will need to be redone when flooded - as will the the FDH. The problem is redoing a copper punch down is a lot easier than cutting a fibre connector off and splicing a new one on (and some are pretty waterproof). Also a copper punch down will limp on for a while - but that is not the point - fibre is great but it needs to be done properly by people familiar with the local conditions (and know what they are doing), not this mad one size fits all rollout.

As far as telephonre are concerned - for those who have bothered to read the latest NBN Co product descriptions - you will NOT be able to get a telephone (Voip) unless you take a data line as well - (fineprint pg 25). Why do you think it is so important that Telstra pull out of the market with their copper - and how do you think NBN Co are so certain they will get 70% takeup ? I pity the samll ISP's or those companies whose business plan is based on a VoIP telephone service only.

Come on guys - start thinking for yourselves and stop swallowing the garbage put out in the press releases.
sydneyla
Jan 8, 2011 9:53 AM
adavion thank you for your easily understood and simplistic encapsulation of the NBN system. Considering the lack of comprehension of the technicalities of the NBN by the average Australian it is my intention to request Senator Conroy to have your explanation copied and delivered to all Australians so as they may gain the knowledge you presented in such a easily understood format.
X_Selectar
Jan 10, 2011 8:45 AM
Sure there's some great discussion about some advanced systems, hardware, plans; do we know what we are going to do with NBN. It will be very "state of the Art", but technology growth is growing at such a rate a whole new communication system could likely be defined in the next year, the engineers, technicians, and scientists are discovering new science daily. We cannot argue any points, we don't actually know what technology will exist in five years.

Alas, the madness will continue.
Someone come up with a use for these high speeds when my 1 Mb/s ADSL 2 is just dandy at home.
umbria
Jan 10, 2011 11:34 AM
@Rossyduck, thanks for spotting my faux pas, which I will put down to holiday brain. Still, can anyone offer technical advice about whether a suitably designed FDH cabinet in a flood-prone locality would be more or less reliable/costly to restore to service than a copper POTS cabinet? Certainly FTTN cabinets would be a right-off if their electric DSLAMs were flooded with mud and water. I have seen pictures of several different NBN cabinet designs, and it would seem uncharacteristic of NBNCo not to have distinct solutions available for all the common situations such as this. Anyone with technical knowledge here?

Far from dreading the NBN, I think that VoIP-only operators such as MyNetFone and Pennytel will be delighted to grow the market for their services beyond the current ADSL base. Many ADSL users have shied away from VoIP due to historic call quality and fear of the unknown, forgoing massive savings in their call costs. But the great customer service these specialists have developed should deliver them increased market share of a huge market, comprising nearly every premise with its uncongested fibre or 12 Mbps wireless connection.

There will be a huge take-up of only 12 Mbps in the early years, in my view, but by around mid-build (2015-16) there will be a massive increase in the proportion opting for services requiring the higher speeds. At a certain point NBNCo will generate revenue faster than forecast. As David says in the article (remember the article?) this is about conservative planning by NBNCo, to their great credit.
umbria
Jan 10, 2011 11:57 AM
@adavion, you raise some good points (inter alia) including the non-loss of any hospital beds from a project that will so obviously pay for itself, and therefore not take funds from the health budget.

But the issue of total network bandwidth needs to be considered in the light of the major component of future bandwidth demand, which will be television. The native multicast design of the NBN requires only one backhaul stream of each popular channel to each node, regardless of the number of downstream premises viewing it. More TV channels do mean more backhaul streams, but you can imagine that more than half of the peak prime time video data volume is going to come from three or four channels at most.

The effect is of course is that any need to increase backhaul capacity will be postponed, and revenue will be well established by then to fund it when needed.

There is a similar benefit regarding the timing of new funds for hospital beds and medical staff for monitoring post-op and elderly patients from country towns when many can more easily have a twice-daily face-to-face chat and blood pressure check from their homes on a take-home video device with a BP strap. Even if 10% of beds are freed up by remote monitoring, this will help our hospitals enormously. Expect a lucrative new high-tech local industry to spring up to design, supply and maintain these devices throughout regional and urban Australia.
adavion
Feb 23, 2011 11:49 AM
@umbria - quick one ....

First in a streaming system every channel has the same data requirement irrespective of who is watching it-

But more interestingly.....

The existence of multicast actually increases network traffic. The reason being is that MC makes it economical to transmit video traffic as part of the peak network load whereas alternative P2P streaming would be too expensive and thus is generally shaped to ensure it does not exceed the spare capacity of the network of whoever is streaming it (i.e. a zero marginal cost).

eg updates to Foxtel on Demand is multicast overnight and locally stored on the Foxtel STB.


Interestingly the number of channels in a Pay-TV network (and hence bandwidth required) increases in proportion with the number of their subscribers. There is an optimum average subscribers per channel number which generates the adequate return on investment IRR = WACC. As subscriber numbers increase, the Pay-TV provider adds new channels until the network reaches the point where adding new channels becomes unviable (i.e. IRR < WACC).

If the network offered fewer channels it would end up with abnormally high returns encouraging a redistribution of value throughout the media value chain or the introduction of more competition.
umbria
Feb 24, 2011 4:00 PM
@adavion, thanks for these observations, but I don't think they change things for the worse, on the contrary.

What is the effect of IPTV, live or on-demand?

Extreme peak demand for free-to-air viewing happens during grand finals of sporting events, major breaking news stories and major television events. Daily peak demand happens when a handful of primetime evening programs are on air.

On-demand video will increase, of course, but a lot of folks are simply going to buffer live programs using their PVR and watch them later from the local copy.

Every household watching or recording ABC1 can be served from the same backhaul copy to their node. If this was not the case, then every additional viewing household would cause the date of a backhaul crisis to be brought forward.

Until the 1980s there were two channels in a regional area. Then there were five channels, and with digital we already have fifteen free-to-air channels. Add Foxtel and you get another thirty. So right now a node must be able to support at least 40 or 50 simultaneous live IPTV feeds, some of them high-definition, but they can be delivered to tens of thousands of connected premises in the suburb. This is the way that multicast postpones the need for backhaul upgrades. It has no effect on random video, which is not to say that Foxtel couldn't potentially schedule simultaneous delivery of some types of requested video.

umbria
Feb 24, 2011 4:14 PM
If one high-definition TV channel can be delivered within 10 Mbps of bandwidth (is this correct?), then 100 live channels need only 1 Gbps of backhaul, thanks to multicast.
adavion
Apr 8, 2011 6:40 PM
@Umbria….

Hmmm yes. From a pure technical perspective you are correct that multicast reduces bandwidth requirements. However when the commercial dynamics are overlaid, the existence of multicast increases network traffic.

Re the Question for IPTV: This will only work as a remotely served on demand service when there is no QoS guarantee, because retail subscribers cannot afford to pay for the on-demand service if it requires its own backhaul allocation.

Consider the cost of using the NBN to do video on demand: You will need around a 12Mbps CVC (costing $24) and you will need CVC capabilities. CVC is priced at $20 per Mbps, so you need $240 divided by the contention ratio.

In a point to point VOD world the contention ratio must be no greater than 1 divided by (1 – percentage of people who do will not be watching TV in peak viewing periods). i.e. If 80% of people watch TV during AFL grand-final time then the maximum contention ratio is 1 / 0.8 to 1 = 1.25 to 1. Therefore the amount of CVC capacity for each subscriber is 9.6Mbps which costs $192. I.e. with the single assumption that in peak viewing periods 80% of people will watch TV, the wholesale cost of NBN connections will be $216 per month. This would typically equate to around $436 per month retail. Furthermore you will also need to buy Tariff Class 3 instead of Tariff Class 4 data on the AVC- which I have not factored in. Bottom line VOD using base load backhaul capacity is too expensive to be commercially viable. Therefore no one will build backhaul to run VOD traffic.

When you allow for multicast, all of a sudden it subscription TV becomes economical ($2.50 per Mbps CVC shared among all users plus $5.00 per subscriber based on NBN wholesale pricing) to do multicast using base load network capacity. Therefore the introduction of multicast to a network increases overall traffic volumes.

Your analysis of PVR and buffering is interesting. As far as buffering goes what you are describing is not true VOD because it is not “on demand”. The point of buffering is to ensure that the RSP’s backhaul limitations are not exceeded.

What you seem to be contemplating is a hypothetical subscription television service running VOD with the media head ends co-located with each FAN site. This is the most bandwidth efficient way of running STV- and it makes a lot of sense. To offer a pure VOD system for the 70,000 subscribers in a CSA where they can access a 25Mbps HD3DTV stream you will need 1.75Tbps. This is well within practical boundaries for a co-located system.

However it is doubtful whether Mike Quigley will allow co-located equipment in a FAN site. That means getting NBN Co to install 200 (one for each FAN) Terabit sized AVCs to link the 200 distributed media head ends of each STV operator to each CSA’s FAN. [Note assuming no scale in pricing here implies this will cost 2000 x $150 = $300,000 per month in AVC costs per FSA to serve 70,000 people- which is around $4.20 per premise, so this is commercially acceptable provided you have significant scale.]

Therefore assuming NBN Co agrees, and then agrees to allow the routing and agrees not to charge for bandwidth flowing through the FAN itself, then and only then does multicast mean a reduction in video volumes.

However if distributed VOD makes sense, no one will be multicasting because it is an inferior service. Therefore multicasting will not reduce traffic volumes because there will not really be any multicasting.

On the other hand, STV providers may be horrified by the costs involved in laying 200 terrabit cables and connecting them to NBN FANs, and decide that it is more simplistic to use the multicast circuits. In this case the multicasting, makes the case for streaming video and hence increases overall traffic.

Therefore in all commercially practical scenarios multicasting data can only increase traffic volume.

[One thing though: Message to Mike Quigley – can you PLEASE NOT invent new acronyms for existing pieces of infrastructure / concepts each time you put out a press release]
Comments have been disabled for this article.
 
 
 
Top Stories
Australian miners send drones to work
In-depth: Unmanned aerial vehicles in the resources sector.
 
The New Zealand telco problem
Opinion: Could Telstra save Kiwi telcos?
 
IT price probe to 'name and shame' gougers
Industry ducking the issue, committee claims.
 
David Havyatt
Sign up to receive iTnews email bulletins
   FOLLOW US...

Latest VideosSee all videos »

Latest Comments
Polls
Should the Government enact new legislation to protect copyright holders in the digital age?

   |   View results
Yes
  19%
 
No
  81%
TOTAL VOTES: 511

Vote