The Inside Track: How will G.Fast actually work on the NBN?


This article was originally published for Delimiter Members only. In late September 2016, Delimiter ceased publishing new articles. Because of a number of operational and other factors associated with this decision, we subsequently withdrew membership articles from publication. If you would like to see a copy of this article, please contact Delimiter directly with your request. Requests by Delimiter Members will be granted. We will consider all other requests on their merits.


  1. is that the NBN company will physically return to premises where it has already deployed FTTB and plonk a small ‘node’ box on top of or into the existing fibre termination unit in the basement of apartment blocks.

    Renai, not sure who told you this but it doesnt ring true to me – the upgrade AIUI is a matter of replacing the relevant Line cards. eg the middle card in this small AL ISAM:

  2. How will be delivered to FTTB subscribers is not my major concern, assuming NBNco haven’t been lying to everyone and that most FTTB users will get 100/40Mbps thanks to the magic of vectoring.

    I am more interested in how or even will be delivered to those on 5-1000m loops from a FTTN cabinet.

    X/ uses frequencies in the 100-500MHz range compared to the 17MHz of VDSL2 currently being used. It is a fundamental fact of the underlying physics, if you want to use higher frequencies you must shorten the loop.

    So how does X/ help those on shitty 5-1000m loops? It is not enough to just upgrade the line card in the FTTN DSLAM if it is running over the same loop.

    An upgrade to X/ for those on long loops seems to imply yet another round of cable pulling, pit remediation and civil works and stills leaves at least one more round of labour intensive upgrades to make the final step to FTTP.

    Every xDSL based upgrade to copper to date and going forward will involve shortening the copper loop that services the end user, as we get closer to the user that gets more expensive to do. This is why I have advocated for FTTP from the beginning, it does it all at once and before labour costs are inflated.

    • No one is seriously entertaining on anything less than 100m loops. A rollout would involve deploying a PON to a pit/pole/etc mounted 1-5 port DSLAM.

      I call it ‘Fibre-To-The-Letterbox’ for this reason. The primary advantage over FTTP is not having to install an ONT in or outside the premises.

      If the user base is already on VDSL, in theory, you can cut the users over to the pit DSLAM and then send them a new CPE in the mail rather than scheduling appointments etc.

  3. As mentioned previously, the 100 to 500MHz spectrum is chuck full of essential licensed services.

    Trying to re-use that spectrum on un-shielded twisted cables is absolutely guaranteed to cause interference to those services and in turn to suffer interference from them.

    The cause of interference to (and from) twisted pairs is imbalance in the pair. The idea of Vectoring is to inject an unbalanced signal into each pair, so as to exactly compensate for the imbalance. Hopefully this will prevent interference at the far end, but by definition will cause additional interference to (and from) outside services. Note that Vectoring cancels cross-talk from within the cable, but it does not (and cannot) cancel interference from outside.

    Experience suggests that will be allowed an initial trial by the ACMA, but then the minefield of resulting radio interference will result in the trial being abruptly terminated.

    It has every indication of being a repeat performance of the Broadband over Power Line farce.

    • Note that the Spectrum used should be 100KHz to 106 MHz (with possible extension to 212 MHz).

      The fact that they have agreed to notch out the FM Broadcast band demonstrates the interference potential. Next there will be a battle about notches for Aviation, Amateur, Business radio, DAB, etc, etc.

      A search on “G.Fast radio interference” shows up a number of early posts expressing concern.

  4. G.Fast is a solution looking for a problem (that we currently don’t have).

    Namely, a short copper pair, in good condition that is part of a vectored group of cables; it doesn’t make sense for the majority of greenfield and brownfield scenarios that NBNco is dealing with, given the maximal line length being considered.

    It is actually optimal for short run, managed pairs; that is, MDU and other dense housing where fibre tails are problematic from and access/ construction point of view, and where it’s plausible that a large number of pairs can be terminated to the same equipment, to leverage vectoring; so this is where I’d expect it to turn up, if anywhere at all.

    Co-existence with VDSL is problematic at best (if VDSL2 is classed above G.Fast by ACMA, then it’s virtually a pointless exercise); it would really need to be one or the other, which means a rethink of where each node sits.

    So, I can see this being something for MDUs and potentially Greenfields; but scarce anywhere else. It’s also hard to believe there’s no political drive behind this, as some kind of “proof” that copper can be the end game.

    Don’t get me wrong, the speed potential is pretty crazy. Crazy good if robust and reliable. But it’s barely ratified and has a very different set of requirements over the existing VDSL based FTTN being built.

    It’s a bit of smoke and mirrors, at this point.

    • They wouldn’t deploy it in Greenfields, because nobody in their right mind is laying new copper in 2015.

      Other than the NBN management that is ;-)

Comments are closed.