Hacker News Reader

Comment by PaywallBuster | Hacker News Reader

optic fiber may be one of the zero interest rates phenomena coupled with gov subsidies betting on the future

At the end of the day there was plenty of juice to squeeze from existing infra

>At the end of the day there was plenty of juice to squeeze from existing infra

That's the argument that POTS was making 20-30 years ago. But now it's just rotting in the ground in our area. In fact, just today I was driving by a telecom pedestal that some traffic had hit and was wondering if anybody even notices anymore. The next intersection there was one that was barely in better shape.

~100 years ago they built the telephone systems. ~50 years ago they built the cable systems. Sure, they can squeeze some more out of the cable, until they can't, at which point we're going to wish we had invested in some more modern infrastructure.

Coax is still going to run out of bandwidth eventually, so fiber lines run will not be wasted, but yeah, it is further out than most laypeople realized, going by other threads about coax, there's this popular conception that coax is super crappy, in reality it's a pretty good RF waveguide, unless it gets pierced or the connector is poorly crimped, but that also applies to fiber

Tbh using DOCSIS4.0 is probably a mistake and a dead end technology. Virgin Media in the UK said the cost of upgrading to 4.0 from their 3.1 network was only half that of actually laying full fibre to all their covered premises, so they are instead rolling out FTTH everywhere instead. 4.0 requires significant new physical network components deployed close to end users so it's not a cheap or easy upgrade.

Plus it's unlikely there will be much more speed after 4.0 given it is using very sophisticated modulation techniques and using nearly the entire spectrum of the coax cable.

> Tbh using DOCSIS4.0 is probably a mistake and a dead end technology. Virgin Media in the UK said the cost of upgrading to 4.0 from their 3.1 network was only half that of actually laying full fibre to all their covered premises, so they are instead rolling out FTTH everywhere instead.

Let's put this into proper perspective.

The state of Texas alone is ~2.8x larger than the entire UK.

Comcast FY22 Residential Broadband generated ~5x more revenue (spot FX; about the same circa end of CY22) and serves ~5x more customers [1; pp. 42-43] than all of Virgin Media O2 Fixed-Line over the same period [2; pp. 5-6].

To think what might make infra capex business sense for Virgin Media in the UK would be asymptotically transferable to Comcast in the US as a broadcast telecom with 6 decades of history and at the prevailing cost of capital strikes me as quite naive.

[1] https://www.sec.gov/Archives/edgar/data/1166691/000116669123...

[2] https://news.virginmediao2.co.uk/wp-content/uploads/2023/02/...

Who cares about the SIZE of Texas? Allow multiple fiber Carriers to operate in the state and build a statewide common infrastructure. Presuming there can be only one is limited thinking.

Consider Internet infrastructure to be a core state & national economic mainstay. Hold your governor and state legislators accountable.

Pressure your Federal representatives to do the same.

The USA has lost the will to all aspects of infrastructure, including power, high speed rail, next generation highways, bridges etc forward.

They’ll provide jobs for companies building the tech and for deployment.

First off, the article is about Comcast and what they're pushing on largely existing infra, while my criticism was directed towards a remark which struck me as uncritically dismissing the value prop of deployed tech just because it may make business sense to a much smaller private company with dramatically different capital requirements operating over a much smaller area in a completely different country; your entire remark is unconstructive and entirely irrelevant in this context.

To humor your rhetorical question on why size matters: When you're a publicly traded company operating in 40 states of varying population densities with over $55 billion in P&E assets and $93 billion in long-term debt (whose annual interest servicing requirement alone would consume all of Virgin Media O2 Fixed-Line revenue) on balance sheet[1] predominantly structured around copper infra as a legacy broadcast telecom with 60 years of history, you might start exercising critical restraint as an entitled consumer who thinks scrapping decades of capex investment for the latest shiny can just magically happen at the flip of some imaginary political switch and without substantial burden to the public in which it was intended to serve. Size matters because sustainably maintaining P&E depreciation while servicing a proverbial fuck ton of debt are real business expenses accountable to a market with waffling sentiments.

Comcast already has ~186k employees on payroll today, ~17% of which are in the UK/EU. Did you even give an ounce of consideration to how those existing jobs would be impacted in the face of imminent disruption?

Do you really think TAM will substantially increase just because FTTH is being offered at a fundamentally much higher minimum entry price for basic service? The FCC is motioning towards broadband for everyone, not just the latest bleeding-edge-nice-to-have-shiny for those with excess disposable income who can afford it.

Who do you think will really be paying the billions in capex debt for new shiny to replace that which still has meaningful life ahead of it just to stay afloat? You're fully entitled to be an irrational consumer who trades in a perfectly operable yet underwater vehicle to finance a new EV at high-as-a-kite interest rates, but I'd expect a lot more sustainability/value common sense from telecom executives with real skin in the game.

Building the required infra on debt is one thing, operating and maintaining it an entirely different beast; read Robichaux on Malone for a historical clue. Small ISPs are already crying a river[2] over BEAD grant capital accessibility because for any proposed project requiring $X to build out, of which 75% will be covered by federal grant dollars, just demonstrating the capacity to have 46% skin in the game is apparently too arduous. As much as I'd like to see small players step in to meaningfully challenge large incumbents, what I really don't want are hundreds of flake LLCs grifting the Treasury while pretending to deliver when the grass is green then conveniently jumping ship just when the going gets tough. As a consumer, real world viability as a sustainable business is just as important as the basic service to be provided.

Furthermore, when the FCC needs to step in and drop the hammer on the multitude of clowns who will try and inevitably fail, it'll be us taxpayers paying those administrative costs, further setting the stage for Comcast et al to acquire shiny assets and regain abandoned market share at steep clearing discount.

[1] https://www.sec.gov/Archives/edgar/data/1166691/000116669123...

[2] https://arstechnica.com/tech-policy/2023/09/big-isps-will-do...

I'm well aware that Texas is bigger than the UK; and I'm also aware that Comcast is far bigger than VM.

I also said 'probably' a mistake, perhaps Comcast is right in doing this as maybe they have split the CMTSs aggressively and the cost of 4.0 is much lower for them than VM.

The other issue that VM is facing is energy costs which are much higher in the UK than the US. Coax requires a lot more power than (XS)GPON; which is pretty much all passive and therefore cheap to run.

Regardless, the long term future is going to be FTTH, not coax.

Average age of coax line in TX might be newer too, since the state has had a lot of new property development in recent decades.

I have a feeling that quality of the existing coax might be a factor in how successful or costly speed upgrades can be, since we're now squeezing the last bit of bandwidth out of them.

Cables that are out of spec but still mostly working at lower bandwidths are going to be a problem since higher bandwidth is more sensitive to the quality of the cable.

Some lines will have to be replaced, which will increase costs.

The issue is rarely the cables - coax run in the last 50 years is essentially identical to coax run now. It had to be to support all those TV channels!

Connectors do sometimes corrode, rodents do occasionally eat things, and they do need to upgrade some of their splitting/serving equipment though.

Yes exactly, to the last part, and by out of spec I meant because of wear and tear.

That's interesting. I thought that, too, but recently I read that there's about another order of magnitude of bandwidth increase to squeeze out of coax yet. I didn't save the link though.

Sorry, I meant relative to fibre. I'm sure there will be DOCSIS 4.1 getting another 20% out of the plant, and various other small incremental improvements. Whereas really there is no reason you couldn't have terabit-PON systems.

Yes, coax (and DOCSIS) is far better than a telephone wire (and DSL).

by time you go n+0 with all the nodes to enable DOCSIS 4.0, you probably could just do FTTH for nearly the same cost.

FTTH opex is also lower after 5 years.

I know nothing of the space, could you go into the opex a bit? I naively assume that once the wire is in the ground, all of the maintenance is in powering routers + staff + recovery when a backhoe inadvertently cuts a line.

How do coax and fiber differentiate themselves on long-term costs?

Coax has powered equipment in the field. Power costs money (opex). The powered equipment sometimes needs replacing, probably much more often than GPON (opex to roll the truck, capex maybe to replace broken equipment)

Precisely

The issue was always mostly about monopoly power and not technical limitation. A hypothetical second coax network would achieve the same end result, but there’s no reason not to go with fiber for new installations today.

Reply on news.ycombinator.com