You can tell from the title and my previous writing I’m not much of a fan of Hybrid Fibre-Coax (HFC), any more. Two decades ago it was one of the most interesting, promising technologies to deal with people’s insatiable desire for bandwidth. We geeks all wanted a 10Mbps cable connection, it was light-years ahead of the 56K modems we had to deal with.
I remember the first friend I had who got Optus cable. We used to download files just to watch them come down at 1MB/s+. This was the future for us.
Over a decade since I first saw Optus’ HFC in action, I’m left wondering why we’re focusing on a technology that has been surpassed by many technologies since, notably DSL and fibre technologies.
The idea that a government would purchase, upgrade, and extend the HFC network is ludicrous. The cable alone will cost far more than fibre, and you have to install boosters, fibre nodes, and the like.
If we look at the review’s “findings”, even without taking into account the technological problems, we can’t go past the redaction of prices. We cannot believe the “$100 per premises” figure earlier in the document when the actual upgrade cost and the actual cost per premises has been redacted.
One can only assume this cost is far higher than the Liberal party would wish us to see. This is all without including ANY cost of the actual network. In fact, the review clearly states that only renegotiation would be required to get this HFC farce off the ground.
One would assume that Foxtel aren’t just going to shut their pay TV network down to accommodate this new and unimproved NBN for free. Something that is a real possibility when we look at the review’s claims that they will attempt to open up another 70+ channels for data on the cable network.
Now Simon Hackett, a man I had admired for a long time, has come out swinging seeing as he’s now a part of this NBN farce, posting an article that is a wild step in the opposite direction to much of what he has stated to this point.
Simon is under the impression that 4-7Mbps on an 120Mbps service equates to “low contention-ratio” broadband. That’s closer to a 30:1 contention ratio, one of the highest hardware contention ratios I’ve seen for broadband.
Meanwhile, GPON (Gigabit Passive Optical Network), even with the 32 fibre splitters being fully utilised, has a contention ratio of 1.33:1 (the NBN only utilises ~19 lines from the split, so it’s a 0.79:1 contention ratio). Even FTTN doesn’t offer such ridiculous contention ratios as HFC, sitting closer to a 3:1 for 50Mbps, and 6:1 for 100Mbps, if you can get those speeds.
I’d actually like to quote Simon himself at this point:
The demand for broadband has to magically plateau on a permanent basis at 2x-3x the current demand that we see in the market.
All the CISCO lovely VNI graphs have to be wrong in the future in a way that they’ve been right in the past.
- Simon Hackett, CommsDay Sydney 2013
The HFC network assumption is even worse: it assumes that the network has already hit this magic plateau.
What’s troubling is that peak speeds aren’t actually being quoted. There are examples of 3:1 ratios, minimum speeds (the 4-7Mbps figure), but no actual figures on what speeds would be offered through the HFC network.
A bigger problem is that they intend to move from 64-QAM to 256-QAM AND increase the number of channels available for HFC. Both these actions alone present some large risks, and high costs to NBN Co.
Firstly, QAM doesn’t tend to play nice once pushed into the higher numbers. Essentially, the larger your QAM constellation, the more susceptible to radio frequency interference (RFI) the system is. While HFC’s cable does actually have great shielding, the weak points in the network (notably the Optical Node and the HFC modem) that can suffer greatly from interference, especially if not shielded well.
There’s many frequency ranges that overlap with HFC, and any attempt to move from DOCSIS 3.0 (the current standard used by Telstra) to DOCSIS 3.1 pushes the top end frequencies deeper into the 3G/4G sphere (topping out at 1750Mhz). There’s cases of LTE phones distorting cable TV and causing drop outs for users.
Even if specific frequencies are discarded to mitigate the risk of wireless devices interfering, it will not remove all the risks as described by this Motorola study.
One thing that strikes me as odd is Simon’s rationale for deploying HFC hinges on the cost of lead-ins:
Where existing lead-ins are up to the task, then it is far more cost efficient and more sensible to use them, than it is to replace them.
It’s funny you say that Simon, as the review itself clearly shows that over 1/3 of those that the government intends to leave languishing on HFC don’t even have an HFC lead-in:
This is all without mentioning the network redesign that will be required to do this “node splitting”. It’s essentially where a cable run will be split into 4, so fibre will need to be run deeper into the HFC network.
The key with HFC networks is that they were designed to cover as many premises as possible between nodes. The current split is around 1 node to every 600 premises on the Telstra network, although much higher in areas that have Multi-Dwelling Units (MDUs). Redesigning this network will make deploying FTTN look like child’s play.
Much like FTTN there are power considerations, booster placement, and having to haul all that fibre past all those houses, install all those lead-ins, yet tell them that hauling fibre is too expensive and the most expensive part of the network is the lead-in (not really, but hey, Simon says).
Looking at the review, it makes some big assumptions about unproven technologies & techniques that even the cable providers Rousselot’s team use as examples aren’t using them.
In fact, every example in that list are dumping HFC the same way Optus is (maybe “was” now) because it’s just not worth the cost to maintain & upgrade HFC. Even as far back as 2004 cable providers were realising the maintenance costs of HFC were far beyond that of fibre technologies when looking at the long term (ITU article).
Even the review concedes that HFC’s speeds will be unacceptable by the time the project is finished, yet somehow this is acceptable? I think not.
All this review is designed to do is line Rousselot’s mates at Telstra’s pockets. There is no logical reason to buy a network that the current owners don’t even want people signing up on (have you tried to sign up for Bigpond Cable or Optus Cable recently?), that one even had plans to fully dismantle by 2018 with Singtel-Optus’ Australian CEO citing:
This decision enables Optus to focus on delivering better service standards and more choice for customers through the NBN.
So the people who run one of the HFC networks know it’s crap & just want people to move to FTTP.
Sorry Simon, but spending billions to buy a network no one wants, and then spend billions more getting said network to offer speeds in line with the current averages doesn’t sound smart. All the misleading claims in the world can’t remove the reality that investing in a dilapidated HFC network is beyond stupid.
Trust this government to make FTTN actually seem attractive. Not through improvements in speed, but through having an even worse option.
I’ll leave you with a quote from a once wise man:
Inequity is something I think sucks in a world where the whole point of bothering with this stuff is to blow the speed problem away.
- Simon Hackett, CommsDay Sydney 2013