The biggest problem with these grants is that it is tax payer money that is then used to build private non open network monopolies.
Easiest solution would be what we have in Switzerland. If you want that government money, the fiber you lay needs to be accessible to any competitor otherwise you aren't getting any of it.
This is the only reason we have 25gbit symmetrical available for only around $70/month [1] because the fiber has to be available for anyone to use.
Recently the largest phone provider here tried to lay p2mp instead of p2p which would result in requiring splitters setting the max speed and features to what they define giving them a competitive advantage over anyone else. Luckily Init7 took them to court and they are now back peddling. Sadly that phone provider wasted an enormous amount of tax payers money and now need to undo the damage they did.
> Easiest solution would be what we have in Switzerland. If you want that government money, the fiber you lay needs to be accessible to any competitor otherwise you aren't getting any of it.
How would you square that against one of the core objectives of targeting rural, unserved/underserved communities? E.g. consider Iowa[1], a single state which is ~3.5x larger in land (144,660 km^2), ~63% smaller in population (3.2 million), and ~90% lower density (22/km^2) compared to Switzerland...never mind that there are roughly a dozen other states in similar or more challenging situations.
Suppose a small business establishes infra in a way that must be "accessible to any competitor". After laying the ground work and establishing a customer base, what's to stop an asymmetrically well-capitalized competitor from stepping in with the intent of swallowing a relatively small loss to operationally bleed out a much smaller incumbent with no margin? E.g. if your local addressable market were half the population of Iowa with service aggressively priced at what Switzerland customers pay, we'd still be talking about a notional revenue cap that's ~1% of what whales like AT&T and Verizon with growing fiber segments generate annually.
I can imagine how outcomes might work out favorably for consumers when the addressable market is a small, wealthy sovereign nation with significant population density...much less when considering rural communities in fly-over states that the rest of the country has a tendency to neglect yet whose economic contributions remain significant.
Those competitors would still have to pay rent to the company that originally laid the fiber, for prices regulated by the government. The open access regulation ensures that all the additional services like
* whether it's easy to get to talk to a human on support
* whether there is TV streaming packages
* whether CGNAT is used or dedicated ipv4's are exposed
* whether there is ipv6 support
* upload and download speeds
* how much the service gets worse in the peak hours
etc, would be entirely decoupled from who invested into the last mile infrastructure. It would be prohibitively expensive (and wasteful) to require competitors to lay a second set of fiber cables next to the existing ones, because in the end it's all paid by the customers.
It's accessible to any competitor, not FREE to any competitor. The business that builds the infrastructure gets to charge a reasonable fee to use it.
I'm from Sweden which has built lots of rural fiber in the northern parts of the country. One setup is that the local power company builds an open-access fiber network and lets any ISP buy into using it, with the power company not even offering ISP services themselves. The power company already has lots of know-how, infrastructure and right-of-way that they can leverage to build and maintain a network of cables easily, but they aren't interested in running a bunch of customer-facing internet services.
> what's to stop an asymmetrically well-capitalized competitor from stepping in with the intent of swallowing a relatively small loss to operationally bleed out a much smaller incumbent with no margin?
What's to stop them anyway? Out here in rural Canada, the local ISP built out a fibre network to all the farms a decade ago.
Then, the government thought rural areas didn't have internet access, so they decided to step in and put up funding to another ISP to build a second fibre line to those farms which is currently in the process of being trenched.
It shall be interesting to see how that plays out once the second ISP lights up.
I live in a small town in a flyover area. I have 300/10 cable (up might be more, I don't bump into it). It doesn't do me much good to have access to 1 gig symmetric that costs more.
It sounds like you're not part of the addressable market[1][2]:
> BEAD prioritizes unserved locations that have no internet access or that only have access under 25/3 Mbps and underserved locations only have access under 100/20 Mbps.
Since congress never repealed the Telecom act of 1996, I wonder if the FCC could decide to reinstate it, as long as they applied it evenly instead of exempting cable like they did before; leading telcos to complain and it to not be enforced at all.
Of course, enforcement on telcos was pretty iffy too. While initial DSL rollouts were subject to linesharing, expanded rollouts were designed to hit the exception so linesharing wasn't required. Additionally, it was pretty common for telcos to have retail prices less than the price for linesharing, so competitive carriers couldn't have price parity without losing money even before considering operating costs.
If we get new legislation, we really need something with real separation; the company managing last mile should be separate from the company providing IP services. A last mile provider would be on the hook to get (ethernet) packets between my house and somewhere nearby where the IP service provider picks them up. This could be at a central exchange building, a local internet exchange if one exists, cable company headend locations, or somewhere mutually agreeable otherwise. But, there's not really any appetite for this kind of thing, so we're kind of stuck. sigh
P2MP is fine -- you can use EPON/GPON/XGSPON/10G-EPON that are designed for shared optical networks with passive splitters or even run DWDM over those splitters to provide P2P circuits over those same infrastructure (PLC splitters really don't care what you're doing with the optical signal so long as it meets the optical signal budget). Cutting the cost of feeder fibres is worthwhile unless you happen to have infinite money available.
The addional cost per connection was estimated at only around $50.
With the splitter any 3rd party will always be limited with what the ISP has installed or decides to allow. Who decides who gets which part of the optical budget? That is considered "market manipulation" here.
Depends on the market. If you're in a high density area the cost of feeder fibres is inconsequential. If you're in a low density rural market, additional feeder fibres are expensive. Smaller market players can't get the 30 year amortization that governments can which makes cost matter.
The optical signal budget in a Passive Optical Network (PON) is a function of the split ratio (commonly 1:32 => 2^5 => 5 * 3dB = ~15dB) and distance involved (typically 0.35 dB/km), which both are a function of the network design. Passive optical networks are typically designed around a 20 km distance with a 1:32 (sometimes 1:64 for 10G PON variants) splitter ratio. You can trade splitter ratio for distance if needed. An 80km DWDM optic is tight at the extreme distance, but 120km is pretty close to the 1:32 / 20km link budget. My point is that it is viable to overlay DWDM on a network designed for PON.
1:32 split gpon in theory can do upto 20KM but in practice with a typical number of splices, reasonable optical budget and so on you are looking at maybe 15KM. 1:64 xgspon is really only good for about 6km.
Cost differences between PON and ptp are minimal in fibre plant typically £5-40; the money you save on spine fibre is mostly lost on splitters and space for them. The real savings in PON comes from head end kit (OLT) and resulting power and space savings. In practice with a mix of 32 and 64 splits port and power costs at the head end is about a tenth of p2p. Of course you make sacrifices in return for that.
Regarding your point on DWDM, it's technically possible on good quality PONs with very high power optics but it would have to be with MACSec or similar as you are giving the signal to everyone on the PON to receive! The costs of this plus optics, dwdm mux/demux and so on makes this a frankly laughable proposition.
This is fundamentally the point. You can't really provide business services such as leased lines or mobile tower backhaul over a PON. In this circumstance you'd have to give a PON to each service provider. This architecture doesn't really economically work for layer 1 fibre wholesale access, you'd realistically need to have the wholesale provider run the OLT and ONT and give each wholesaler layer 2 access to each customer. So for example you may be stuck with GPON (on openreach) like we are in the UK.
This debate was had at the time between ofcom and Openreach here and layer 2 access was deemed acceptable. Several years later, they had to force them to provide duct and pole access and in some cases dark fibre (for business and backhaul use in areas with no competitors). The former was mostly because openreach were too slow to rollout even thier PON solution, so we now have a situation in some areas where as many as half a dozen ISPs have installed thier own fibre all the way into each persons flat/house.
Would it have been better to let a contract for p2p fibre install in rural areas rather than subsidising openreach on a PON and then in some cases have a commercial fibre provider build over that unsubsidized. Yes, no doubt.
A 1x32 splitter costs somewhere between $20-70 (unless you pay exorbitant Corning prices).
32 core (well, 36 core) cables are cheap enough, but that's not what you need for a pure point to point network. Let's say you've got a local convergence point with 576 fibres facing the distribution network. With 1x32 splitters you need 18 feeder fibres. You can get a 24 core cable for $350/km.
Now, let's say you're building a fully point to point network. You now need a 576 core cable for your feeder in the access network. That's going to cost north of $6000/km (that's a wild guess -- sorry, I don't have accurate numbers for high count cables as I haven't bought any in the past few years).
If your access network is 10 km, that's the difference between $3,500 in cables vs $60,000 in cables. Plus you're adding a good $1,500-3,000/km in cost for larger splice closures and increased splicing labour.
Where it really hurts is when you start getting close to your Central Office (CO). With P2P networks, you need more larger cables in the first few kilometers closest to the CO, which is often the most congested part of a network. For brownfield builds, that might require directional boring to install new conduit in a town to bring the new cables into the CO. That can get really expensive.
Designing fibre networks is a game of whack-a-mole. Spending less on one component drives up costs on other components. Sure a P2P network means you can use cheaper SFPs, but the outside plant is more expensive. There's good reason that most carriers are building Passive Optical Networks -- P2P just isn't economically viable in all scenarios.
These maths are quite theoretical! In reality in a 1:32 PON you'll typically use somewhere between 14 and 30 ports (because typically you want to leave at least one spare port). In rural areas the lower side of that is more likely as your splitter placement will be governed by how far apart properties are.
864f cable which is the biggest you'd realistically blow is about 11mm diameter, and costs around £9/metre. If in the PON situation you're blowing say 72f for £1/metre then your costs are honestly not that much higher, considering you don't need to have lots of splitters all over the place (often you have actual passive splitter cabinets, as putting them in chambers is not always possible) you may need a bigger splice joint but not really much bigger because you're now using ribbon fibre. The number of splice operations would be similar too. Obviously this doesn't scale super well, but we are discussing rural subsidised fibre here, not low cost competitive urban deployments. The likelihood is 864f is more than enough for a whole village, or even several. If not add a second.
Yes, but you don't plan your network based on actual capacity usage, you plan for a fully loaded network that can serve 100% of dwellings passed along with some spare capacity to serve additional buildings over the expected lifespan of the network (30+ years if you're building an expensive incumbent network, 5-10 years if you're trying to minimize cost). Anything else means you could end up turning customers (revenue) away.
As a software engineer, I would use 25gbit daily, particularly during EU and US work hours (my team crosses continents).
I'd download those yuuuuuge deployments in seconds instead of minutes.
And the scary thing is... I'd be super tempted to actually get it if it were offered here. I already have 1gbit fiber and the hardware for 10gbit. I wouldn't need to upgrade much to utilize it.
Well, I would if the servers I'm downloading from could provide at that speed. I'm pretty sure they're connected at 10gbit for a lot of the same reasons.
No server is going to give you 25 gbit for yourself. Or routes are going to be congested. I have 1 gbit and almost all servers don't have that available for me.
I'd happily switch to it tomorrow if I could get it close to that price. I'd probably even eventually get around to getting switches and NICs that'd let me actually see the speed.
Are there benchmarks websites that show average fixed broadband performance for Switzerland? Switzerland is not that in the rankings on Speedtest for example.
I am not aware of such a site. Customer to customer I have 25 gbits, uplink I believe is 100gbits shared between each 48 port which can have modules between 1 and 25gbit.
Init7 is the only ISP I know here that offers this speed. Most people are happy with 1gbit or less and there are also still many areas that do not have fiber plus the over 600k illegal p2mp connections that can't be used. Also many of the others have an insane oversubscribe factor.
The proponents of "getting a subsidy to pay for internet" of course were the entrenched players. More than likely, we'll see ISPs double dip: raise their prices to what they think the market is willing to pay (which they do now) AND collect the government subsidy, "raising prices to optimum cockbag levels": https://www.youtube.com/watch?v=0ilMx7k7mso&t=63s
The idea that we need to give grants to utilities that are ostensibly better as private institutions since private companies allegedly "drive innovation" is so nonsensical on it's face. All it's led to is a cadre of entrenched corporations that do not give the slightest fuck about their quality of service or uptime, and competitive pricing is a fucking joke in the American broadband market.
I have not heard any arguments in favor of this system that hold water in my entire life and it's only getting more and more true as consumers get screwed harder and harder with zero solutions on offer either by these mythical "competitors" or the Government that acts as defacto enforcer of their ability to choke customers out of more and more money for worse and worse service year over year.
> The idea that we need to give grants to utilities that are ostensibly better as private institutions since private companies allegedly "drive innovation" is so nonsensical on it's face.
If the idea was we needed to give grants to drive innovation, something was missed. There's not really an unmet need for innovation in broadband delivery; the issue is running new wires broadband has costs, and in a lot of places, the ROI on running new wires for broadband doesn't justify doing it. The grants were supposed to induce large providers to do the runs in areas that weren't commercially viable; of course, it's easier to ask for grants and then do nothing with them, which is even better for the bottom line.
Really, we'd be a lot better off if last mile had mandatory line sharing (as the telecom act of 1996 provided), and if there was some cooperation from the major broadband providers to coalesce on some upgradable standard for it. A single fiber to the home built on GPON/10GPON/NG-PON2 could be leveraged by telco, cableco, and others, but since telco and cableco will only use their own lines and won't build unless there's demonstrable ROI, a GPON overbuild by a third network is doomed to poor financials: either the incumbents were right and you won't see the uptake, or the incumbents were wrong, will notice when you overbuild and will update their network and the 3rd network still won't get customer uptake.
Where I live, I only get VDSL2; the cable company comes to the corner of my property but won't serve my house. Municipal fiber is available, but only if I pay to run it above ground a couple miles to the drive, and I've also got to pay for undergrounding from the pole to my house. If I could get a few people together, some of the costs would be shared, but not a lot of people are willing to pay tens of thousands of dollars to get better internet, when they have a serviceable alternative now, and the monthly costs are likely more with muni fiber too. Of course, if the cableco or telco decide to roll out to my house, the installation fees will probably be minimal, because those companies can account for the costs in other ways than a public agency.
> There's not really an unmet need for innovation in broadband delivery; the issue is running new wires broadband has costs, and in a lot of places, the ROI on running new wires for broadband doesn't justify doing it.
This sounds like exactly the sort of thing a private company is going to be terrible at. And even with the grants, it seems they are and have been since I was in preschool.
Yeah, grants don't really seem to work here; my point was more that if you thought they were for innovation, that was way off-base. The grants can be helpful for upstarts, but not when the criteria exclude upstarts.
If the goal is universal service, it really needs to be mandatory, not voluntary, even with incentives (especially poorly administered incentives). But mandatory service is usually connected with a granted monopoly, which isn't really what we want either.
On the other hand, if it takes long enough to get everyone wired, wireless options will become competitive. Between LEO internet services (StarLink, maybe OneWeb and friends) and terrestrial 5G-NR, I think it's quite likely that few places without current wired broadband will ever get it.
>I have not heard any arguments in favor of this system that hold water in my entire life
One defense is that the old AT&T system was really expensive and slowed progress. Deregulation was a big win for consumers.
I agree that well regulated monopoly ISPs could deliver better service. But I'm deeply skeptical that, in practice, we'd manage to actually regulate the ISPs properly.
Our regulated power companies have shitty, underfunded grids that create massive forest fires.
The phrase you're looking for is Cronie Capitalism. It's "capitalism" (note the air quotes, please) with the gov putting it's thumb on the scale to benefit the few. The joke is the "free markets" are better narrative and Cronie Capitalism is anything but free markets.
The municipality should own the network and lease bandwidth to private providers. I'm sure private providers would be willing to help lay cable if they got a fixed fee for the work done along with a period of slightly reduced lease rates and first dibs on a fairly sizable chunk of the pipe's leases. The fixed rate would cover work expenses with a little profit and the cherry lease deal would pay out better than a bond and be fungible so it'd be easy to secure credit on that agreement.
The same thing happened here in the province of Ontario. The government handed out roughly $2.5 billion in subsidies through a reverse auction, but not a single smaller provider was able to participate. The lots were too big (my local area had a reserve bid of $51 million), and the bonding requirements were excessive (want to bid on that $51 million lot? Better be bonded for $51 million even if your planned bid is going to be a fraction of that). I don't know of any carriers that bond their projects -- if the builder doesn't make progress in a month, then they don't get paid.
It's pretty sad that governments don't hand out telecom subsidies in a competitively neutral manner. Funding Make Ready costs for repairing pre-existing conditions and addressing lack of Spare Capacity on existing pole lines would make it easier for all carriers large and small to participate in the market. Beyond that, obligating pole owners to pre-engineer telecom attachments would really help. The region I live in is primarily served with aerial communications strand on poles as there are too many rocks, hills and trees to make wireless or burial viable for many of the unserved.
Easiest solution would be what we have in Switzerland. If you want that government money, the fiber you lay needs to be accessible to any competitor otherwise you aren't getting any of it.
This is the only reason we have 25gbit symmetrical available for only around $70/month [1] because the fiber has to be available for anyone to use.
Recently the largest phone provider here tried to lay p2mp instead of p2p which would result in requiring splitters setting the max speed and features to what they define giving them a competitive advantage over anyone else. Luckily Init7 took them to court and they are now back peddling. Sadly that phone provider wasted an enormous amount of tax payers money and now need to undo the damage they did.
[1] https://www.init7.net/en/internet/fiber7/