Australia's $44B Broadband Network May Settle For Fiber Near the Home 229
Garabito writes "In April 2009, Australia's then prime minister, Kevin Rudd, dropped a bombshell on the press and the global technology community: His social democrat Labor administration was going to deliver broadband Internet to every single resident of Australia. It was an audacious goal, not least of all because Australia is one of the most sparsely populated countries on Earth. ... So now, after three years of planning and construction, during which workers connected some 210 000 premises (out of an anticipated 13.2 million), Australia's visionary and trailblazing initiative is at a crossroads. The new government plans to deploy fiber only to the premises of new housing developments. For the remaining homes and businesses — about 71 percent — it will bring fiber only as far as curbside cabinets, called nodes. Existing copper-wire pairs will cover the so-called last mile to individual buildings."
Re:Pragmatic choice (Score:5, Informative)
Well it was never going to be FTTH ~everywhere~. The original plan proposed by the Labor government was for every town with more than 1000 people to have FTTH, with the remainder being served with either fixed wireless, or for the most remote 1% or so, satellite (which is already available of course, but the plan included a significant upgrade of satellite speeds and capacity). Doing the calculations, it essentially meant 93% of the population would get FTTH.
The Liberal government from the outset said that if they got elected, they'd scale back the FTTH and rely mostly on FTTN/VDSL for existing developed areas (though, still supporting FTTH for new greenfields development, since if you have to lay cable anyway it may as well be fibre). As you say, that's probably fast enough for most purposes provided you can keep copper line lengths down to a few hundred metres at most.
The criticisms of this revised plan, broadly speaking, are that:
1. Much of the existing copper is in bad condition and would need to be replaced anyway anyway to deliver decent VDSL speeds and reliability. Telstra, responsible for managing the copper network, has publicly stated that they consider the copper network at end of life.
2. The Liberals' plan, compared to the original Labor plan, would only result in cost savings of 20-30%, yet deliver an outcome that is a lot more than 20-30% worse (in terms of speeds, reliability and future capacity for growth and upgrades).
Re:Don't they have an fiber to the node cable netw (Score:3, Informative)
No. Most people don't have cable, but instead have ADSL over copper phone lines from the interchange to the home. Pay TV is not ubiquitous, and AFAIK is mostly served via satellite. I live in a fairly typical suburb and the interchange is a few kilometres away, so max download speed is around 4-5 Kb/s.
Re:Don't they have an fiber to the node cable netw (Score:5, Informative)
When you say 'cable', are you referring to cable as in US-style cable TV (and internet, using DOCSIS)?
If so, then no, most areas of Australia do not have this. Subscription TV is delivered by satellite in virtually all areas of Australia, save for small sections of urban Sydney, Melbourne and Brisbane. Far more cost effective for such a big and sparsely settled continent. So the cable footprint would be lucky to cover 5 or 10% of the population.
Currently most people in Australia get their internet via ye olde copper phone line using ADSL2+ (which can provide up to 24 Mbps if you have a short line, but degrades rapidly and can barely push a few Mbps at distances of 4-6 km, depending on the quality and gauge of line).
FTTN rollout would thus require that nodes be built, branching out from or replacing the current telephone exchanges/central offices (where lines currently terminate) so that they would be no further than a few hundred metres from any given house, and leverage the existing phone lines as much as possible to cover the remaining distance. You can push 50-100 Mbps using VDSL2 over these kind of distances. But only if the lines are in good condition (which they aren't, in many cases).
It should also be pointed out that most newer areas (built in the last 10 years or so) already have fibre right to the door, and also that some parts of the original FTTH NBN network have already been completed (I have some friends that are already on it, at 100 Mbps). But the rollout is still only 10% complete at most.
Re:Not an issue, provided... (Score:5, Informative)
Actually it IS an issue. The previous government was planning to spend $37bn to provide FTTH which gives a future proof network and guarantees a starting speed of 100Mbps. By that I mean the next logical upgrade is a simple change of gear either end of the fibre making the network future proof. They were anticipating that most households would be able to cheaply upgrade to 1Gbps internet in the future.
Compared to that the current government wants to spend $20bn and provide FTTN at 25Mbps which I don't consider much of an upgrade from the current 25Mbps.
The previous government's plan was to spend double the amount to upgrade the internet for most Australians, and the current government is still spending a fortune for what will not be an upgrade for people in major cities, is in fact slower than the two major telcos current cable networks, gives benefit of a fast cabled connection to a few coastal towns, and then sticks the of rural Australian on either high latency satellite, or an overly congested wireless link.
As for upgrading the last mile if there's money to be made, you don't really understand the way these networks here work. We live in a country where some of the installation of the last mile was so cheap that people couldn't get more than one phone line to their house. That's right they split the 2pair phone line between 2 houses. It's a country where the last mile of copper is rapidly corroding due to cheap maintenance over the last 20 years. Even in major city centres its somewhat accepted in areas that your internet will drop out when it rains. Oh better yet the last mile is owned by one company.
I am still wondering how the coalition promoted the former Telstra CEO who absolutely destroyed the value of Telstra, who accepted that fines from the Ombudsman were a cost of doing business, to the CEO of NBN Co. It's almost like they deliberately want this to fail. I can think of better things to do with $20bn than waste it on nothing.
Re:What's the speed limit of copper? (Score:4, Informative)
VDSL2 is great in the lab but in the real world the speed numbers up and down can drop off.
http://www.zdnet.com/nbn-co-cant-guarantee-libs-50mbps-speed-promise-report-7000023901/ [zdnet.com]
"....only realistically be offered two guaranteed speeds: 12Mbps (with 1Mbps uploads) and 25Mbps (with 5Mbps uploads)."
Re:Not an issue, provided... (Score:4, Informative)
Note: this post is from a UK perspective, things may vary a little arround the world.
By the 1980s, we had developed mathematics and modems that could use the same lines to get up to about 33.6kbps at 3,429 baud.
And then things more or less stopped there. There was one more marginal speed increase (56K) but they had pretty much hit fundamental limits of the phone system. Pushing speeds further required bypassing parts of the phone network.
ISDN BRI delivered slightly better speeds in the 90s but the way it was priced (if you wanted a 128k connection you had to pay for two phone calls in addition to the ISDN line itself costing more than twice what an analog phone line did, AIUI most unmetered dialup packages allowed single channel ISDN but not dual channel ISDN) made it an expensive option. ADSL turned up in the early 2000s but again it was initially expensive.
ADSL gradually improved through the 00s first with the providers getting more confident and taking the artificial limits off and then by the providers moving to ADSL2. However while speeds improved so did the gap between the haves and the have nots. Those close to the phone exchange could get 20mpbs, those stuck a long way from the exchange got less than 1mbps and we have pretty much hit the limit of what phone cables can carry over long distances even with advanced modulation techniques.
In any case, odds are, whatever we put in the ground today, in 20 years we'll be able to do more with it than we can today.
The problems with mixed fiber/dsl systems don't really have anything to do with the fiber that is being put in the ground.
1: it still relies on that old phone wiring for the last hop. There are a few tricks we can pull but we have pretty much hit the limits of what those cables can carry over those distances. You still have the "cable length lottery" except now it's distance from the point of fiber to copper transition to the house rather than distance from the phone exchange to your house.
2: having all that infrastructure spread out like that makes it very difficult to do incremental upgrades. When ADSL was introduced they could start by putting one DSLAM in a phone exchange and patching the subscribers to it, when one DSLAM filled up they could add another. It didn't matter that only a few percent of customers were taking DSL intitially because the phone exchange was large. On the other hand there were places in the UK that had their POTs and ISDN delivered over an early fiber to the cabinet system and these were among the last to get ADSL because it wasn't worth putting a DSLAM in a cabinet for a handful of subscribers. So even if there was a system that could get a slight improvement over the current VDSL gear rolling it out would be very expensive.
The only real way to substantially improve a "partial fiber" system (fiber to the cabinet, fiber to the distribution point etc) is to push the fiber closer to the subscriber but each time you do that your infrastructure ends up even more spread out. Eventually you get to the point that you may as well just take the fiber all the way.
On the other hand with a fiber to the home system all you have to upgrade to deliver faster speeds is the consumer premises equipment and the exchange equipment. All the outdoor infrastructure can remain the same. Plus current fiber to the home equipment has a much wider margin over current needs than the VDSL that is being deployed in current fiber to the cabient system.
Deploying fiber to the cabinet now means in a few years time either internet speeds will stagnate again or a fiber to the home project will be needed anyway making the fiber to the cabinet equiment redundant.