My previous article on the “Destruction of the NBN” was widely read and a few questions were asked in regards to some of the questions or statements I raised. Many commenters noted that FTTN isn’t all that bad – and in comparison to the alternative (being nothing) this is true. The problem is that, unlike FTTP, FTTN performance is based greatly on your personal circumstances. It also relies heavily on the last mile copper being well maintained over time, the node being within sufficient distance and that power is continually supplied to service the connection.
I had initially hoped to follow up my original article with a rebuttal by nbn™, who did initially make an attempt to reach out to us but did not reply to any follow up emails. Instead, I thought I’d break down the various parts of the “Mix” and the pros and cons that relate to each aspect. I will also take into account the longevity of each option and the upgrade paths available, as well as the percentage of the population slated to receive each technology. I’ll rank them from top to bottom in terms of speed, quality and longevity.
Note: The percentages are off mainly because the Mix isn’t an exact science. FTTP/HFC may see more punters if the copper situation is worse than feared.
RANK 1: Fibre to The Home (A.k.a FTTP, Fibre to the Premises)
FTTP is hands down the fastest, most flexible, cheapest to maintain and future proofed technology in the mix. Fibre to the premises involves, at its most basic state, feeding a fibre optic cable directly from the exchange (or nbn™ POP – Point of Presence) to your home. Once there, it hooks into an PCD (Premises Connection Device), which is basically a box on the outside of your house, before snaking inside to be plugged into your ONT (Optical Network Terminal). The ONT (Or NTD as it is also known) is what hooks up to your Wi-Fi router to provide a connection to your house. The ONT will also have ports to allow Phone and may in future also provide TV or Foxtel connections.
The major benefit of FTTP is its ubiquity. There is a staggering amount of available bandwidth on a single fibre pair, thus the number of services and their breath is very high. Other private fibre networks, such as Opticomm, already send TV and Foxtel services via fibre to their customers, alongside Phone and Internet services as required. Switching on higher bandwidth services, such as gigabit or multi-gigabit internet services, both up and down, is more of a backhaul and software change than anything else.
Fibre, being a single thread of glass, is also extremely wearable. It is not affected by water, exposure to air or the ravages of time. It requires no repeaters or external power facilities to keep it active – a single fibre run can be over thousands of kilometres with zero signal degradation as long as the correct optics are utilised at both ends. It’s easy and environmentally sustainable to make and very cheap. Power is only used to “burst” the light signal from one end to the other.
Fibre, once installed, wouldn’t require direct maintenance (outside of damage to conduit or deliberate sabotage) for decades. It could easily serve the needs of the user for at least 30 years, possibly longer, and only, really, if someone developers a technology that is faster than light. To illustrate my point – the core of those thick bundles of cable that run between nations to provide inter-country/continent connections are the exact same that run down your street and into your home.
Roughly 25-28% of the population are slated to receive it, those lucky sods.
RANK 2: Hybrid Fibre-Coaxial (A.k.a HFC, Cable)
Hybrid Fibre-Coaxial is not very well understood. Many attribute it to Foxtel, and the unfinished rollout of the 1990s. At its core, HFC is a very underutilised technology with a staggering amount of money behind its continuing development. HFC works in a similar manner to FTTN, except that instead of copper running from the node to your house, it’s coaxial cable. The node converts the light signal into an electrical “RF” signal, which heads across the coaxial cable (with the help of a few repeaters to ensure a strong signal).
HFC uses a software standard known as “DOCSIS” in order to provide a high speed internet connection across the network. The latest, version 3.1, theoretically allows download speeds of up to 10Gbit/s and 1Gbit/s up, which is comparable with some of the highest commercially purchasable speeds in the world. But there’s a few catches – unlike Fibre or FTTN, HFC is a shared medium, meaning there is a pool of data at each node that is split between its users. If there is an adequate investment in equipment and backhaul, then this shouldn’t be noticeable to the end user. But as users of Optus’s atrocious HFC network can attest to, this can cause dramatic drops in performance.
However, there are many benefits here over FTTN. Thanks to repeaters, HFC nodes can be much larger and further away, meaning in many cases a node could be in a building suburbs away rather than on the side of the road. Coaxial cables do use copper, but this copper core is tightly packed in by both a dielectric insulator (to reduce the issue of noise) as well as metal and plastic shielding. As a result, newer cables and equipment can pack a lot into a very wide radio spectrum without losing anything.
The technology is old, but the software and hardware that are being introduced are very new. DOCSIS 3.1 was only ratified 2 years ago, and has been running on the Telstra network since 2014. New CMTS equipment from ARRIS has been purchased by nbn™ and will be setup as the new backbone of the network. A large bulk of the aging Optus gear and nodes will also be replaced to bring it up to spec as well as whatever parts of the Telstra network that haven’t been upgraded yet. Once in place, there isn’t really a lot stopping nbn™ from flicking on Gigabit connections for all, as long as they can keep the data pool topped up.
But it’s not all roses. HFC networks are expensive – to power, to maintain and to keep ahead of the curve. When new customers come online, the network needs to be closely monitored and modified to keep contention ratios (basically who gets how much bandwidth) fair, so no one person is at full speed while someone else is not. This was such a problem in the mid-nineties in cable obsessed America that competing DSL networks would advertise their “one to one” guaranteed data allocation. Once would assume the software has caught up now to manage this however.
Then there’s the issue of future proofing. DOCSIS is a remarkable unicorn in that it keeps producing great results, but there’s only so much you can squeeze out of RF. I wouldn’t be surprised if DOCSIS 3.1 is the last major iteration and any further upgrades are just pushing the bar a little longer in order to stay competitive until the move to fibre is completely unavoidable. Still, for the purposes of the nbn™ over the next decade or so, 3.1 is not a bad second choice.
Roughly 34% of the population are likely to get this technology, mainly in those areas that either cable networks already pass. Due to the cost of new rollouts (almost everything else is cheaper) it’s very unlikely anything outside of a few streets in HFC heavy areas will be rolled out.
RANK 3: Fibre to The Node/Curb/Basement (A.k.a FTTN, FFTC)
FTTN is probably the most well-known and equally, the most contentious part of the mix. It famously relies on the existing, ageing, copper lines installed by Telstra over the past 100 years and will be the cause of over 20,000 new street side cabinets in order to house new equipment. But how does it actually work?
Just a quick aside:
FTTN/FTTC/FTTB (Fibre To The Curb/Basement is just another fancy way of saying FTTN basically) involves running Fibre as close as possible to the household, but stopping short – usually between 100-200m – in order to build a cabinet that converts the fibre into a bunch of various DSL connections to each house via the existing wiring (in 99% of cases, copper). It’s much cheaper than fibre mainly due, in most cases, labour. Putting fibre in the ground and hooking it up to houses is actually where most of the cost profile occurs. In apartment buildings, there is occasionally Ethernet available, which is better, but its normally copper.
FTTN already exists (sort of) in a lot of areas that didn’t have spare ports or room in the exchange – Telstra famously built nodes known as RIMs (Remote Integrated Multiplexer) which were designed, primarily, to save money running copper all the way from the exchange. The major issue was that Telstra didn’t bother to build them to support high speed internet and in many cases, speed was capped at dialup rates until they were replaced by new nodes called CMUXs (Customer Multiplexer) which supported basic ADSL connections, usually capped at 8Mbit/s. Again, cost.
The brand spanking new nodes will probably not look all that different, and its likely many existing ones will just be upgraded. The difference will be the equipment inside will be new and rated to support G.Fast – nbn™’s white knight. G.Fast is a brand spanking new protocol, (its barely out of the lab, nbn™ will be one of the first networks to deploy it) that uses a different method to cook the signal to avoid being affected by noise.
See, the copper that in the ground most of the time is largely unshielded. It loses power and is susceptible to interference from other radio signals as it runs, and thanks to a lack of insulation and repeaters, if it has to run over any sort of distance the amount of bandwidth it can carry lessens. The reason your ADSL2+ connection sucks is because you probably live between 3-5km from your telephone exchange. If you lived next door, you’d be maxing out your speeds. So the aim with DSL is to remove that distance. This is why nodes are important – fibre has no distance degradation, so it really comes down to putting nodes in the middle of house clusters.
G.Fast fixes one of the two massive problems with ADSL – it removes the issues around “noise” via a number of clever techniques. As a result, it can carry a much higher data load without worrying about interference, even though it still needs to be transmitted within a short distance. As a result, the 24Mbit/s maximum of ADSL2+ jumps to a fairly impressive 1Gbit. Realistically, however, the maximum drops to between 150Mbit/s and 500Mbit/s, based on the idea that the node is going to be within about 100-250m.
So this begs the question, why is nbn™ only committing to 50Mbp/s? For two reasons. The current standard in FTTN rollouts is a technology known as VDSL2, which suffers all of the same issues as ADSL2+. Since G.Fast isn’t likely to be rolled out until 2017, according to nbn™, most early access punters will initially be stuck on substandard speeds of between 30 and 100Mbit/s until this occurs.
Going forward, it’s still possible that some clever mob will come along and squeeze even more out of your humble, unshielded copper cable, but it’s not especially likely. Most of the smart money is being dumped into fibre and wireless technologies, while DOCSIS is almost entirely funded by cable companies trying to squeeze that last little dollar out of their depreciating assets.
Roughly 40% will be getting some kind of Fibre to the Node/Curb/Basement/Windmill.
RANK 4: Fixed Wireless
Fixed Wireless is a strange beast. On one hand, it’s probably one of the cheapest ways to rollout broadband available. It involves strapping a bunch of specialised wireless radios to a high tower/pole and then attaching an outdoor antenna to the roof of your house. This antenna feeds down into a router and bam, you’ve got yourself an internet connection. On the other, it relies on technologies that require hard to acquire wireless spectrum to work properly.
There are a number of ways to distribute Fixed Wireless connections but nbn™ is going with 4G-LTE, which is very similar to the connection your smart phone uses. The pluses are that it’s a very reliable, powerful and versatile connection – unlike Satellite, it’s not affected by poor weather and does not interfere with other signals or frequencies.
The minus is that, unfortunately, its maximum throughput is quite low – only 25-50Mbp/s down and 5-20Mbp/s up. This is mainly due to the restrictions of wireless technology and spectrum allocation – basically, only so much data can whittle its way through the air, and all the networks need to share it. As a result, that speed isn’t likely to change in the near future. Unlike the other 3 technologies, software isn’t really able to do much to alleviate this. Spectrum is spectrum. This is also why everyone is carrying on about fixed line connections – wires can carry a lot more bandwidth.
The plus is that most areas receiving this connection are likely to have extraordinarily poor connections already – rural and regional areas that can get mobile phone signals but not landlines. This connection will be much cheaper, more reliable and enable things like streaming video both up and down which would have been expensive and difficult before.
In fairness to nbn™, Fixed Wireless was part of the original brief and almost all of the cons are outweighed by the fact it offers a solution many have been crying out for decades to receive.
Roughly 5% of the population will be connected via Fixed Wireless.
RANK 5: Satellite
Last, but not least, we have Satellite. If you are remote enough that none of the about 4 options are suitable, then this is your only option. Thankfully, it’s not completely terrible. Satellite internet works on what is called “2-way”, in that your Satellite receiver not only receives but broadcasts back on the same system. Earlier satellite internet systems would actually upload on a separate landline dialup modem, if you can imagine it.
As you would expect, beaming between 3 satellites (base station, space, your house) takes time, which means most online games are out of the question. But it does provide up to 25Mbit/s down and 5Mbit/s up. It’s also the only part of the network with its own name – Sky Muster™, named after the 1st of 2 satellites nbn™ launched in October. It’s also the only part of the network that can provide access to offshore islands and territories like Norfolk and Christmas Island.
Long term, it’s kind of up in the air. Satellites generally don’t have much scope to be upgraded, so it’s likely when they reach the point of obsolesce you’d imagine nbn™ would send up another one. Right?
Again, Satellites were part of the original rollout, and are an essential part of the network. The speeds many of us scoff at now, are almost unthinkable for many businesses and residents in very remote areas. It could genuinely overhaul communications in many outposts, some of which barely have satellite phones.
The Mix could arguably be worse. That’s about all I can say. The drop from 93% to 25% for FTTP is genuinely saddening, and regardless of how hard nbn™ are working to make HFC and FTTN palatable, it’s really quite disappointing that the government aren’t willing to lick their wounds and at the very least double the number of households scheduled for FTTP.
On the plus side, the noises coming out of nbn™ in regards to their HFC offering are really quite good. DOCSIS 3.1 is a proven and very stable platform, and there have been numerous assurances, that they are upgrading the entire network, not parts of it, in order to serve those very high speeds they tout are on the radar. It’s not unlikely that at least 2/3’s of the country could be accessing Gigabit speeds within the next 3 years.
The jury is out on FTTN. G.Fast is just too experimental to make any calls on, and the likely launch platform is almost certainly going to be VDSL2. How long it takes to iron out the bugs and niggles is for someone else (or me, possibly) to look at in 2018. But it is nice to see they are attempting to shoot the moon – there are some risk takers still hanging around inside the company.