NBN finally admits something like defeat - and not just with the HFC rollout.
Who could have thunk it? After years of commentary by tech journalists, engineers and general sceptics of the MTM, NBN Co has finally admitted that HFC is not ready for prime time. In a wonderfully spun press conference and subsequent press release, the company basically admitted the one million-plus rollout of their HFC offering has been pretty much substandard. Taking to Whirlpool shows the wide extent of the problems – high attenuation, faulty hardware, massively delayed installs, unstable connections, woeful speeds. The other problem, which was not publicised but fed to me via a source close to the matter, is that the projected cost of HFC, which put it below other options, exploded due to the massive gaps in lead-in throughout the original Telstra network.
This is once again another symptom of the problem with changing course so rapidly and abruptly in the space of a single year, with engineers developing first-draft ideas with no testing outside a few small labs and employees in the field. One of the larger issues with HFC is that it uses a system that is traditionally locked to a single provider. What NBN does is effectively makes “itself” that provider, then creates a virtual “tunnel” that allows other providers to supply their own services over that connection. NBN was the first ISP in the world to roll this idea out on a large scale and it obviously suffers from a lot of teething issues.
Then there is the case of the lack of lead-in. While Telstra’s network grid was wide across the major capitals, there were enormous gaps where the larger rollout stopped after Telstra and Optus initially ran out of puff back in the early 2000s. So, while almost every suburb that existed back then was touched in some way, including the building of nodes and backhaul, many streets weren’t cabled and many houses on cabled streets weren’t lead in. NBN are now finding that the “millions” they found inside a theoretical HFC frame don’t have a single scrap of equipment. This means, ironically, digging up front lawns, running underground and above ground coax, and putting in a host of repeaters, nodes and all the other stuff that was supposed to save money.
This is once again another symptom of the problem with changing course so rapidly and abruptly in the space of a single year, with engineers developing first-draft ideas with no testing outside a few small labs and employees in the field.
The rest of the “quality issue” mainly relates to the problem of relying on decades-old infrastructure, most of which hadn’t been checked or touched since it was installed. Simply dropping in a new modem without checking the quality of the line or if the node is up to spec because, frankly, there isn’t any time for it, has effectively left a litany of destruction in the rollout’s wake. The next months will be NBN engineers frankly trying to fix bugs in their software, actually properly test the hardware, send field guys out to the worst performing areas to diagnose and essentially do all of the things they should have done in the first 9 months.
This is not all NBN’s fault – I know for a fact that they have exceptional staff and they work damn hard. The issue is that they were handed a tray of poison chalices with only one full of untainted wine. There was always going to be huge issues with FTTN and HFC because they are no longer fit for purpose. I still stand by my words on HFC – when it’s working it will be one of the better options (after FTTP/FTTC) because it has a very well-engineered backend protocol in DOCSIS3.1 and its capacity limits are very high. It’s also still being actively developed and supported in the US and Europe. But it’s rushed rollout has done nothing but waste more money, time, and effort for very little gain.
SCRAP IT ALL
Moving on from HFC – in the past week or so, NBN quietly announced that they were going to scrap the existing AVC/CVC arrangements and replace them with one that should have existed in the first place. If you don’t already know, AVC is effectively the charge for your NBN line and CVC is the added tax on speed that flows over it. The old method was that your ISP would be charged one flat rate AVC cost per connection, then a stack of CVC based on how much throughput they (the ISP) deemed each connection required. The changes, which are apparently coming into play next year, will contain an “effective” amount of CVC within the AVC charge, negating a lot of the RSP/ISP complaints that both charges erased any semblance of profit they might make.
This doesn’t solve the CVC problem, but for most customers it will remove the main impediment they have in regard to peak speed issues. NBN has also announced that it will reduce the cost of CVC to the point of equivalence with the new plans so existing customers can get the same benefits before next year. I applaud them for facing common sense on this front, especially since almost no one with any sense thinks that NBN will make any decent commercial return or that anyone would buy it for the price it cost to make. One of Labor’s bigger mistakes involved courting conservatives with the “build to sell” model, which did nothing but inflate build costs and introduce stupid shit like ROI into a public asset.
It was also a quiet announcement for the same reason – not only does it (rightly) devalue the network and effectively admit defeat, NBN wants to keep these changes strictly within the back channels of the sausage factory. Any inkling that end users will assume these changes will fix all their woes creates a false positive that could come back to bite them.
Generally, however, this is a positive. As is my next point – refunds!
Several NBN RFPs finally admitted that they were setting plans unfit for purpose and have offered refunds before effectively been forced to. Both Telstra and Optus notified customers that had paid for plans that exceeded their sync speeds and offered refunds and contract cancellations. This is a good and bad thing – RSPs will be much less likely to offer plans that sit on the top end of potential speeds if they feel there is a chance they might not able to offer them 24/7. For example, if you have a 100/40 connection it will probably sit around 75-80/30-40 during peak times – this is normal. But Telstra might just sell you a 60/25 connection instead because it’s less risk for them. Scary.
One final note – many people on twitter who already had HFC connections or scheduled orders were concerned the changes affected them. They don’t. Only brand-new orders were suspended, all existing HFC connections will sit on NBN’s Alpha network until they sort everything out. No worries.