The user, who preferred to use his Whirlpool handle, harryxebec, told iTWire that initially he thought the problems with his HFC connection, which he has had for about a year, were due to his old hardware and his operating system: Windows XP.
"I have always had browser lock-ups, pages not loading correctly and broken downloads. It wasn't a fatal problem and downloads always resumed," he said. "I thought that the problem was the old XP machine that I usually use. Use 'em up and wear 'em out I say."
But then when he encountered a problem of a broken download from a non-resuming server, he realised that he would have to resolve the issue.
"That's when I started monitoring the network and discovered the regular zero bytes received impulses as in the screenshot (below) that I provided earlier and is visible on both computers."
In a post on Whirlpool, he said that while observing a single 50Kbps download using the Windows Task Manager, "the bytes received formed a steady spiky wave".
"Increasing the refresh rate of the display showed that each dip was, in fact, a short interval, I estimate as approximately a few to several or more milliseconds, where the bytes received always fell to zero. Each dip occurs about every two seconds like clockwork. Every time a download breaks, it is always during this zero bytes interval."
Later. he posted that he might have found the solution himself. "The problem is called 'impulse noise' and is usually caused by electrical equipment. As the impulses appear every two seconds, I don't think that this will be my case.
"HFC uses the DOCSIS system which sets equalisation parameters every 20 seconds. If an impulse occurs at the same time the parameters are read, the system receives and operates on corrupted data for 20 seconds when there is a good chance that the right parameters will then be sent."
He said the chances of this taking place "several or more milliseconds out of 2000 every 20 seconds means that the dropouts would be intermittent and upredictable".
"It also explains why sometimes Web pages do not load properly, my Web browser freezes (malformed scripts), the delays (occur) in download restarts and why the larger the download, the less likely it is to succeed."
When he first encountered a non-resuming download, harrybexec said he had done what anyone would usually do: "Reset equipment and replaced cables. Two days ago, my ISP asked for a factory reset on my router which did nothing except disable my VoIP for over 24 hours. I suspect that the impulses had something to do with that as well."
A resident of Salisbury North in South Australia, he is still not 100% sure that his conclusions are correct. "At this stage, I cannot be 100% sure that this impulse noise is a factor," he said. "However, I do think that it is. I feel that, at this stage, I have only identified a potential problem that I have not heard of before and that it will have serious implications if it is a fact."
He said this was why he had posted on Whirlpool and also informed the media, "hoping that someone will query NBN Co about it and that tech heads somewhere will investigate further".
Contacted for comment, Robin Eckermann, who led the creation of TransACT and served as its chief architect during the 2000-2003 network rollout, did not have a direct opinion on harrybexec's post, but said that there were two key reasons why HFC had been rejected during the TransACT rollout.
"In creating TransACT, we came under great pressure from HFC champions to implement a HFC network rather than the FTTK/VDSL architecture that we ultimately deployed," he told iTWire.
"There were two key reasons why we rejected the HFC option. The first was concern over the potential for noise ingress (with the coax plant acting as a giant antenna picking up all sorts of RF interference).
"The other was the reality that HFC was originally designed for one-way transmission of TV signals, where everybody receives a common payload and selects the stream in which they're interested. While DOCSIS developments have come along in leaps and bounds in retro-fitting HFC for interactive communications, we felt that the FTTK architecture, inherently designed for interactive communications from the outset, was the safer choice. We never looked back and regretted that choice."
iTWire has also contacted HFC expert and business consultant Dermot Cox and NBN Co for comment as well.
Having spoken to his ISP about the issues, harrybexec is hoping that the issue could be resolved through NBN Co.