To say that everything surrounding the Australian NBN is a political can of worms would be an understatement. The general elections in a few weeks, with Labor trailing in the polls, may spell if not the doom at least a radical trimming of NBN’s ambitions. In this context, I stumbled upon this interesting presentation by Simon Hackett entitled Building a Fiber NBN on a Copper Budget. I don’t know who Simon Hackett is, and if he is politically motivated (it seems like everybody is, looking from the outside), and more importantly perhaps, I haven't done the financial analysis that would allow to assess whether his opening statement is correct.
What's more interesting to me in this context is that he suggests some interesting ways in which a wholesaler can lower his own investment costs. In a nutshell, he offers three paths (not competing, complementary):
- Forget about QoS: basically, he argues that bandwidth trumps QoS, and since fiber offers virtually unlimited bandwidth, the complexities and costs of QoS management and more importantly QoS wholesaling should be avoided. I don't disagree on principle with this one, although the confines of QoS should probably be better defined, but I'm assuming he means "different grades of traffic management".
- Drop PSTN (and multiple VLAN capability): Simon's point is that PSTN is dying and shouldn't be carried over to the NBN. I agree with him 100% on that one although I'm pretty sure the regulator imposed PSTN continuation to the NBN anyway. But honestly, any fiber deployment done today should have two aspects embedded into its DNA: eliminate copper (long term) and eliminate legacy systems. Not sure I'm quite as bullish on eliminating the ability to deliver separate VLANs. Some services, especially around home security, healthcare, etc. are going to require a fully separated path for security reasons, and until that can be done over WDM, VLANs seem to me to be the only way to go.
- Let the ISPs pay for the ONT: that's the most intriguing of Simon's suggestions. Instead of imposing an ONT paid for by the NBN, he says, the set-up should assume that the ISPs will install the ONT since they will want to install some equipment inside the home anyway. On paper this looks very tempting and financially it could represent significant savings for sure, but I'm concerned this simply couldn't work for a simple reason: first, unless the NBN forced a vendor onto the ISPs, which seems unlikely to be approved, I'd have serious misgivings about interoperability of the ONTs. This has been a known issue with GPON for a long time and while various vendors have spoken about furthering interoperability in the standards, letting the ISPs pick their ONT vendors would still be a huge leap of faith.
Still, these are some interesting ideas, and at least it's good that they're being asked. I'm assuming in saying that there is no political calculation behind them. In which I may very well be wrong.
Still, if Simon Hackett can convince Malcolm Trunbull that he can get a fiber network for the cost of a copper one the Coalition should be trumpeting that left, right and center!