The NBN (National Broadband Network)

[fatprop]

http://www.theguardian.com/world/vi...-turnbull-nbn-internet-speed-whiteboard-video

I'm no liberal party fanboy, but Turnbull is one of the few that is emerging from this clusterfuck of a government with his shoes relatively unstained.

It is a pretty good way of explaining the "challenge"

 

[Tex]

The natural argument to Turnbull's point about marginal utility (and it is a good one) relates to future broadband requirements.

Technology changes exponentially; both in its application and the infrastructure requirements to support it. Applying current use to future estimates isn't always accurate.

But irrespective, it's a good video and should have been put out by the government a long time ago.

 

[Pfitzy]

Of course its a challenge. Anything good is.

IF they can set it up in a fashion that allows it to be easily upgradeable (pull my fibre connection through as they reel the copper out) then I'm happy to go with that, provided I can get a speed above 5MBps in the next three years, and something above 50MBps in the next 7.

 

[Pfitzy]

Fuck - good news everyone! The Coalition-funded report into the NBN says we're $16 BILLION better off using the Coalition's FTTN!

http://www.news.com.au/technology/o...government-model/story-fnjwncel-1227038690275

The report found that the FTTP would have a net cost of $35.3 billion, compared to the Coalition’s pared-back model, which would cost $24.9 billion.

BUT IT GETS EVEN BETTER!!!

The analysis — the first on Australia’s largest infrastructure project — also shows the Coalition’s cut-price version would deliver a net economic boost of $18 billion, compared with $2 billion for Labor’s gold-plated version.

Wow. Gee I'm glad the Coalition told us that. SOOOO fucking glad.

 

[Braveheart81]

The best bit was that the report specifically didn't consider whether bandwidth requirements would increase beyond 2023 "because there was no evidence they would".

A friend of mine made quite an eloquent post on the subject (I thought).

My biggest problem with an NBN Cost-Benefit Analysis is the biggest benefits are unimagined, and unimaginable.

The people who are going to imagine, develop and realise the greatest benefit from the NBN are only just being born.

They're going to be raised in a digital, connected world, and come to use that digital connectedness in ways that we can't imagine.

So unless there's a line item for "Unimaginable, world-shaping, concept-changing future developments", there's doesn't seem to be much point.

 

[boyo]

No doubt that Runner will be along soon to tell us that everything is hunky-dory.

 

[fatprop]

The best bit was that the report specifically didn't consider whether bandwidth requirements would increase beyond 2023 "because there was no evidence they would".

A friend of mine made quite an eloquent post on the subject (I thought).

The interesting thing I thought was the newer technologies that rest higher bandwidth over the copper as well.

Whilst it looks lovely for us all to have 1Gb connections, the only real winner "today" would be services like Netflix (who are massive chewers of bandwidth) and whilst there may be some driving benefit to break that Foxtel monopoly, it is a lot of money to dig up every street in the country to lay fibre to do so. It would be cheaper to buy foxtel

Realistically I reckon if you want to have those extreme full fibre connections you will be moving to newly built networked apartments, new housing developments and businesses with the need will move where it is available or paying for your own fibre being laid from the node.

The rest of us will deal with a lessor product (but if lessor is 50gb, that would do me, there is only so much porn I can watch).

It is a compromise, but in the end Governments don't have unlimited budgets and have to make decisions on the cost/benefit.

 

[Braveheart81]

When you're building a piece of infrastructure that should last for between 50 and 100 years, it's crazy to only look at what it is capable of today.

When the copper network was rolled out, computers weren't even a thing, let alone the internet.

 

[fatprop]

When you're building a piece of infrastructure that should last for between 50 and 100 years, it's crazy to only look at what it is capable of today.

When the copper network was rolled out, computers weren't even a thing, let alone the internet.

and yet we worked out how to leverage that network into something that is not perfect, but it works.

If we all lived in the population density of Korea or had no infrastructure the answer would be far simpler, but retro fitting to every house in the country?

 

[Braveheart81]

and yet we worked out how to leverage that network into something that is not perfect, but it works.

If we all lived in the population density of Korea or had no infrastructure the answer would be far simpler, but retro fitting to every house in the country?

The CBA that has been carried out also hasn't taken into account any maintenance on the existing copper network.

What happens if in 20 years time it is decided that the copper network that every house relies on is in such a terrible state that it all needs to be replaced?

I get what you're saying but I think it is a poorly thought through plan that will cost us much more in the long run.

Cloud computing is increasing at a rate of knots and together with a far greater desire to share content, upload speeds are becoming increasingly important yet we're planning a network that does almost nothing to improve upload speeds. That is where I think the biggest folly lies. Whilst not being able to stream ultra HD television will be a luxury we can live without, there will be huge swathes of opportunities that require reasonable upload capacity that we simply won't be able to use.

 

[fatprop]

yeah, but I am a business luddite and cynic who is being pushed kicking and screaming to upgrade from XP (it works and isn't broken)

That cynicism was exacerbated by a few years in IT consulting.

The cloud may be the way of the future today, until the circle returns with CEOs (and their consultants) looking to return to a more "strategic" and "entrepreneurial" model that entails managing resources in house. (They need to keep selling whatever isn't there and CEOs want to change shit, because they need to do something)

 

[Pfitzy]

I personally think this whole Cloud thing is not exactly the magic bullet everyone wants, but I'm thinking about it in terms of ordering data and providing meaningful output. You can't just throw all your files at a NOSQL* store and expect efficiency forever.

However, once most companies get past the "privacy" issues of storing customer data in the cloud, uptake is extremely difficult to stop. Cost is fucking tiny compared to an in-house system during non-peak times, and scalability is guaranteed in a pay-per-use model. Compare this to a pay-for-worst-case model and its easy to see why Cloud is not going away.

After all, why buy shitloads of software licenses for machines that might only experience > 80% usage for an hour, once a month? I can keep my running costs low and have peak times catered for with full transparency.

And the only way its going to be easy to run is if I have loads to reasonably-priced bandwidth to get those results between my office and the Cloud, and then out to my staff if I choose to VPN the access via my own security.


* For those who aren't into it: NOSQL = Not Only SQL. Basically the premise that you can throw any form of data at a cloud store and have it scale up the resources it needs for any subsequent queries.

As a crusty old DB programmer, its pretty clear that when someone actually wants a large query to make sense, a fair bit of heavy lifting has to occur in order to bring through the results people want. Simple searches e.g. "Find this product" are easier.

 

[boyo]

Whilst it looks lovely for us all to have 1Gb connections, the only real winner "today" would be services like Netflix (who are massive chewers of bandwidth) and whilst there may be some driving benefit to break that Foxtel monopoly, it is a lot of money to dig up every street in the country to lay fibre to do so. It would be cheaper to buy foxtel

Realistically I reckon if you want to have those extreme full fibre connections you will be moving to newly built networked apartments, new housing developments and businesses with the need will move where it is available or paying for your own fibre being laid from the node.

Why do you think that you need to "dig up every street in the country to lay fibre"?

Why do you think that you can to "pay for your own fibre being laid from the node"? The big IF is if FOD is available.

 

[fatprop]

Why do you think that you need to "dig up every street in the country to lay fibre"?

Why do you think that you can to "pay for your own fibre being laid from the node"? The big IF is if FOD is available.

1/ How do you get the fibre to your house otherwise?

2/ If I can now pay for ISDN lines etc, I expect to be able to pay for a full fibre connection if isn't provided

 

[boyo]

1/ How do you get the fibre to your house otherwise?

2/ If I can now pay for ISDN lines etc, I expect to be able to pay for a full fibre connection if isn't provided

1. NBNCo paid Telstra handsomely for access to their ducts. The only time that digging would be required is if there is no duct or if it is unremediable.

2. FOD (fibre on demand) doesn't have to be provisioned. If you wish to pay for private fibre, then that is another matter.

To quote Mark Newton from Internode:
“FTTN doesn’t bring FTTP any closer, but it does push it several billion dollars further away….there’s no upgrade path from one to the other. This notion that FTTN is a “stepping stone” to something else is pure fantasy. If an FTTN network is built you’d better like it, because it’ll be around for a long, long time to come.”

There is NO Fibre on Demand/Fibre extension policy!

Turnbull even said so himself on Lateline.

 

[Pfitzy]

I would consider paying for the "last mile" provided I actually owned it.

Which I never could, because it would run on public property through vendor-owned courses. And there is no guarantee of my bandwidth even then, because the aggregation from the Node to the backbone is going to be smaller than a FTTP/H solution.

The assumption for the FTTN situation is that there will be "X" copper connections from the Node, with a maximum capacity of "Y". Therefore, optic fibre to the capacity of "Y + (Z%)" is installed to cater for maximum plus a bit of fat.

When it comes time to retrofit fibre to the premises, then suddenly X copper = (X x 100) fibre, and Y+(Z%) is no longer even close to being able to cater for it. So then the fibre from the Node to the backbone needs radical increase.

Trouble borrowed collects interest.

Do I NEED 100Gb for my house? Nah. Probably not today. Maybe not even tomorrow. I can survive on 5Mbps for today, even though its the primary bottleneck in my otherwise connected house.

But by the time the LNP's solution gets to my less-than-two-year-old neighbourhood, Australia will have not improved its overall speed rating world wide:

2012 - 39th (4.9Mbps)
2013 - 40th (4.3Mbps)
2014 - 42nd (4.6Mbps)

We do very well in terms of mobile connections, with speeds better than Japan. But then, why would they need 4G when they're saturated in WiFi?

 

[Braveheart81]

If the LNP is looking for a three word slogan for the plan, I propose "Yesterday's technology, tomorrow."

 

[Blue]

The CBA that has been carried out also hasn't taken into account any maintenance on the existing copper network.

What happens if in 20 years time it is decided that the copper network that every house relies on is in such a terrible state that it all needs to be replaced?

I get what you're saying but I think it is a poorly thought through plan that will cost us much more in the long run.

Cloud computing is increasing at a rate of knots and together with a far greater desire to share content, upload speeds are becoming increasingly important yet we're planning a network that does almost nothing to improve upload speeds. That is where I think the biggest folly lies. Whilst not being able to stream ultra HD television will be a luxury we can live without, there will be huge swathes of opportunities that require reasonable upload capacity that we simply won't be able to use.

Which ones? Can you list them?

As the cloud concept expands. So does efficient data transfer.

I started my career programming X.25 networks. We implemented the very first dual X.25 network that ran to the customer premise in SA. We were told by the PSTN neanderthal guys that there is no way we'll get it to work over the copper. Right. Two years later ADSL went mainstream on the same two wires.

There are so many other massive latency issues upstream that you can stick in all the fiber you want but you cannot leverage it properly anyway. This is almost never discussed.

As long as there is an upgrade path for copper as and when it is necessary we'll be fine. You don't have to go wall to wall. It's fucking insanity beyond belief, dreamed up by Ruddy on a napkin, followed by a complete reluctance to listen to anyone who actually knew what to do.

Fuck NBN. I wish I had the kind of speed a friend of mine can squeeze out of their Telstra wireless connection a block and a half from my house but I am just outside of the coverage zone by about 100 metres. That would make a big difference to my work although NOTHING to the work I need to do on servers round the world. Dropping fiber right under my arse will make only very marginal difference to that, if at all.

Sure, the copper will die, but I bet you by then the technology will have advanced substantially and we will also be able to define the end user requirement much better. Right now it's all iffy as hell.

Horses for courses. That's what we need.

 

[Pfitzy]

There are so many other massive latency issues upstream that you can stick in all the fiber you want but you cannot leverage it properly anyway. This is almost never discussed.

At least we won't BE the latency issue with fibre in future. The rate of change at the moment, we're going to be chasing 40th spot for years to come.

I always look at it this way: what is the point of Wireless-N device on a smartphone when the bottleneck occurs as soon as it connects to the resource? Theoretical maximums of 300Mbps N are great, but the only time I made it work on a PC was when the router was in the same room - at which point plugging in the Gigabit Cat6 was the preferred option.

At which point, the internet connection was a massive bottleneck anyway.

LTE and LTE-A make mobile connections awesome, but they're never going to be economically sound because you pay through the nose for them. Its not just about speed, remember.

Let's say Turnbull is right, and its an $18bn difference overall if we started today. In the course of 20 years, that's less than a billion dollars a year to put ourselves at the top of the tree and stay there.

I think there is going to need to be a change in expectations for the rural sector. I've lived there, and sorry but you give up your access to certain services when you accept the fresh air and open spaces. I'd find it hard to go back, but as I get older I understand the limits of my mortal existence, so I could put up with that.

 

[boyo]

As long as there is an upgrade path for copper as and when it is necessary we'll be fine. You don't have to go wall to wall. It's fucking insanity beyond belief, dreamed up by Ruddy on a napkin, followed by a complete reluctance to listen to anyone who actually knew what to do.

"Labor’s plan was done on the back of an envelope."

So was Hewlett Packard's and Compaq Computers, then they combined.

So was the US and Australian Constitutions.


However, the work that went into them all, including the NBN, by people with detailed knowledge and expertise moving the "envelope" to reality was massive.