Hacker News Reader

Comment by rayiner | Hacker News Reader

This is a great bill. It creates a slight public subsidy for broadband, but the total amount of subsidy is minimized by tying it to digging that's already going to be done. This should be extended to power line and water pipe construction. No reason you can't put in some dark fiber while you've got someone up there hanging new power line (which I imagine is way more expensive).

Most importantly, it minimizes market distortion. Inatead of making it more expensive to deploy broadband (which is the ironic side effect of many broadband deployment policies), it makes it cheaper for competitors to enter the market.

View on HN

More importantly: we already did this. Remember when they overhauled the highway system in a lot of areas back in the late 80s and through the 90s? While they were digging, they dropped a lot of fiber.

Almost all of the major routes we use today for teleco/internet routing were laid or substantially upgraded then. Almost all the dark fiber laid was laid then.

The hilarious part? Most of the fiber laid could do at least 40gbit, thus the cost of upgrading a lot of the existing backbone could be as easy as just swapping the line cards on either side.

IMO, we may not actually need more fiber (outside of last mile usage, which this bill is directed at), but just better use of existing fiber. If you're wasting fiber on anything less than 10 gigabit, please stop. It costs you next to nothing to upgrade compared to what it cost a decade ago: upgrading to equipment that can handle 10GBase-LR/ER/ZR or OC-192/STM-64 or better is worth every dime.

Literally, if the US government passed a law stating, say, any fiber over 5 miles long must have a line rate of at least 10 gigabit, and heavily subsidize upgrading equipment that is simply too old to handle it, a lot of Internet speed WTFery would go away in rural and suburban areas.

You'd be surprised how many LTE/4G cell phone towers, DOCSIS 3.x headends, or modern ADSL2+/VDSL2/possibly G.Fast DSLAMs do not have 10gbit or greater backbones. It's just absurd in 2017 that we're using stuff that is 90s era at best.

I think the idea is to get fiber into the last mile. Vast parts of the country are still under construction. New subdivisions are being built, and people are moving there from established northeastern/midwestern states. Requiring conduit to be put underground as you build the roads for a new subdivision, for example, would dramatically reduce the costs to deploy fiber in these places.

This would only apply to projects receiving federal funding, which I don't imagine includes most subdivisions. Most likely just interstates, state highways, and maybe county highways. That wouldn't necessarily get fiber through your front door, but it should get it far enough out to provide a viable high speed option for rural households that are currently stuck with dial up or satellite.

There is no reason we couldn't or shouldn't also do the same thing at the state level.

They can make taking federal highway funds contingent on adopting a dig once policy for locally funded projects. This is how the speed limit and drinking age was standardized across the country.

It will be an interesting change, for the federal government to use this power for good rather than evil.

I still find it astounding Americans can't drink until 21.

But if this goes through on a national level it would spark off local legislation as well. This would be a great half-a-loaf for politicians who can't afford to oppose these monopolies.

Roads regularly get resurfaced in places that are already well-established, so it's not just a "new construction" thing.

Resurfacing, especially in rural areas, is exactly that -- the existing road is scored and then paved over. There's not enough room there to insert a conduit without actually digging a channel.

I'm not saying that work shouldn't be done, but it wouldn't be a case of just throwing down a pipe during normal resurfacing work.

wouldnt the equiv be that resurfacing==upgrading optics/lincards... you dont need to replace the actual fiber.

> think the idea is to get fiber into the last mile. Vast parts of the country are still under construction. New subdivisions are being built

Such an interesting subject is this; we have freaking building codes, regulations and requirements - esp re: elec and structure and such...

We need to take a step back and make a freaking INFRA code that requires what the data line requirements to a building should be.

DATA SHOULD BE A UTILITY - just like the sewer, water, etc...

we should stop bitching about FCC this and that - and REQUIRE that construction includes the data plan as well. This will alleviate much of this kvetching.

Please give me your opinion on the above as I respect your perspective.

I think the breakdown should roughly be: developers must wire out to the property line (and if they're building a subdivision, out to the main road). The county should wire anywhere a public road is being built and bring glass to some reasonably convenient place (because otherwise developers would have nothing to connect the other end of the line to). Private companies should handle everything that's not dumb glass.

I think counties are reasonably capable of burying fiber in the ground as they build and repair roads, but that's about all I trust them to do. I live in the DC metro area. Our public sewers dump raw sewage into the Anacostia river every time it rains. The city has an extensive network of old lead pipes that are poisoning kids. The Metro was almost unusable last year because tracks were catching fire. These problems exist all over the country, and are all caused because by prices for services being set too low, because they are set by elected officials and not the market.

Meanwhile, we have had FiOS in MD/VA since 2006-ish. In that time the top speed has gone from 50 mbps to 750 mbps as Verizon upgraded OLTs, backhaul, etc. Yeah, I pay $100 per month for 300/300 service, while my water/sewer bill is so small I don't even know what I pay. But I don't see that as evidence my internet bill should be lower, I see it as evidence that my utility bill is too low.

>These problems exist all over the country, and are all caused because by prices for services being set too low, because they are set by elected officials and not the market.

They exist because the US continues to insist on spreading responsibility out of thousands of local governments instead of having a single set of rules and standards for everyone.

Why are we even building suburban subdivisions and strip malls still? Maybe if we utilized the infrastructure in urban areas (and even existing suburbs) we wouldn't have the need to keep investing in new infrastructure while our old infrastructure crumbles.

> Why are we even building suburban subdivisions and strip malls still

Not everyone wants to live in an apartment building.

All utilities should be owned, funded and built by the Government. Roads, sewers, water mains, power mains, data mains, etc.

A "radical" idea I'm advocating for, is that all of the following is concidered nessesary for the society, and thus should be done by the government, or heavyly regulared (non profit, enduser owned type of companies):

- Roads

- Water

- Sewer

- Community heat

- Healthcare

- Education

- Pension

- Internet connectivity

- Banking (including loans for 1 family houses) and card payment

You just covered a significant fraction of all economic activity (maybe a majority).

Perhaps that was your intent?

Banking as a utility? I love it. Credit unions are already a thing, and we have the National Credit Union Association, it's a small jump to the idea of a National Credit Union.

And we need single payer healthcare yesterday.

Off the top of my head, you're also missing electricity, food, shelter, clothing, and entertainment. Those are more necessary than some of the things you've listed here.

So I guess we'll have the government own or heavily regulate 90% of the economy. Gee, that's historically worked so well in the past... :)

"outside of last mile usage, which this bill is directed at"

How is that? It applies only to federally funded projects. Most homes and businesses are not on federally funded roads. This bill is not a FTTH driver.

FTTH is a type of last mile, FTTH is not the only consumer of last mile fiber.

The future of America is not, and never will be, FTTH. The future of America is FTTF/FTTDP (fiber to the frontage/distribution point) to a street/basement cabinet that houses current generation micro-headends and micro-DSLAMs that are setup to handle the last 500-1000 feet, speaking G.fast and DOCIS 3.x as very small greenfield installations.

Current gen G.fast tops out at 1000/1000 in less than 300ft installations, slows to 100/100 at 1500ft (about a third of a mile) using existing installs of standard 80s and 90s era unshielded phone wire, the numbers go up with brand new copper, and next gen G.fast is aiming at increasing these speeds several fold (talk of 10gbit in under 300ft and 1gbit at much further distances).

Current gen DOCIS 3.1 tops out at 1.3gbit down, 245mbit up, and there are already existing gigabit-on-cable customers getting what they paid for on 90s era coax.

I see absolutely no reason to roll out actual FTTH ala FIOS, and I agree with Google Fiber's decision to drop actual FTTH, and they should be aiming for taking over unmaintained networks that can be rebuilt as modern DOCIS 3.x or G.Fast, or deploying micro-towers doing shorter range LTE-A fixed.

Current gen LTE-A maxes out at somewhere in the neighborhood of at least 10gbit per tower/frequency band, and fixed LTE-A "modems" with outside directional antennas can get gigabit connectivity; next gen LTE (LTE-A Pro, aka 4.9G) at least doubles this.

All of this is useless without Congress pushing for more fiber between towns, or inside large metro areas (which means burying more fiber along federal highways and major state roads).

There is literally no reason why, today, we don't have at least 50/10 to every goddamned home in America. Every goddamned home, no matter how far you are out in the sticks. This is a drain on our economy every second we aren't working towards this actually rather easy to achieve goal.

Fun fact: I don't even have reliable cell phone connectivity at my house. I'm maybe the only person on HN that bothers to maintain a POTS line and use it as my primary phone number.

> There is literally no reason why, today, we don't have at least 50/10 to every goddamned home in America. Every goddamned home, no matter how far you are out in the sticks.

Because the country is huge, and building adequate backhaul to sparsely populated areas is very expensive.

> This is a drain on our economy every second we aren't working towards this actually rather easy to achieve goal.

I find this sort of argument to beg the question. How much would it really increase overall GDP to get at least 25/3 (the current FCC standard) to the 10% of the country that lacks it? Especially considering that everyone can get satellite broadband, which covers the core things people need internet for: job applications, access to online educational resources, electronic payment processing, etc.

We are talking about less than 1/1,000th of our current GDP so no, it's not a question of having a larger GDP. GDP is simply not the problem, it's local monopolies that keep getting funded to fix a problem have no incentive to actually fix the problem.

Remember, we rolled out copper wire to every home with a far smaller GDP, doing the same for fiber is relatively speaking a smaller investment.

I'm asking about the economic benefits to GDP of rolling out broadband to the places that don't yet have it, considering that they can get satellite.

There are a lot of things you can't do on satellite internet right now. VOIP is the most obvious example.

SpaceX's planned satellite network could fix that eventually… if it's ever actually constructed.

OneWeb is in the lead, they actually have funding and launch contracts.

I agree that OneWeb is in the lead, but playing devils advocate: SpaceX has funding (Fidelity/Google) and rockets...

Hughesnet and Exede satellite internet both offer a voip service.

Every home is probably overkill. But every home that is also connected to a road, electricity, water and sewage should also have decent internet. Laying fiber for 50/10 everywhere is expensive, but so is building roads everywhere.

20% of American houses are not connected to a public sewer system while only 10% can't get at least 25/3 broadband.

There is literally no reason why, today, we don't have at least 50/10 to every goddamned home in America.

Where I live in the Texas Hill Country, I get 300MB in the center of town. Outside the town there are a few companies providing wireless Internet connections, but this is limited by topology - the hills for which the region is named tend to get in the way.

I also have a cabin in upstate New York, in the foothills of the Catskills. I can't even get a cell signal there without hiking to the top of the mountain - and even then, a plain voice connection, non-data, is iffy. And I'm doing OK, I know of people up there who don't even have electricity.

I think your intention is good - and I agree we should put in modern fiber.

But only a small percentage of fiber put on long routes uses single wavelengths. Most fiber going more than 5-10km uses DWDM and that is indeed an important place to put in modern fiber.

However, most of the railway routes in 1998 used fiber that supported 32*40G lambdas, so the problem with most of the fiber laid over the last 20 years is not that it can't handle more than 1gbps. Most of the problem is the equipment, not the type of fiber used.

Companies have being running various WDM solutions on backbone links for ages.

I agree with you here. However some of the 10gbit iom/mdas for the Nokia 77x0 bins are $100,000s.

SONET was replaced with DWDM a decade ago. With 100GHz ITU Grid C, you would get 40 channels of up to 10Gbps each. More recently, most long haul providers have been upgrading to coherent using 50GHz spacing with 80 channels of 200Gbps, further broken down into 10Gb and 40Gb TDM or PDM circuits. Long haul density is not the issue.

For a real life example, on a recent quote we received, fiber down the west coast from 2001 6th Ave in Seattle to 624 S Grand Ave in Los Angeles was only 1/3rd the cost of going from 624 S Grand Ave out to Malibu.

The issue is not with long haul density, but with last mile due to restrictive municipal regulations and procedures, and lack of collaboration between different metro networks.

I didn't know about any of this prior to reading your post, so please forgive my naivete. 100% municipal internet access in the US is a game changer for education as education laws like California's CA 61199[1] requires that primary instructional materials be available to 100% of students in both the classroom and at home

In other words, 100% internet deployment or municipal wifi or whatever form it takes, literally changes what is allowable to primary instructional material content in public education in the united states.[2]

[1] http://www.cde.ca.gov/ci/cr/cf/textedcod60119.asp

[2] IANAL and different states have different laws, what I am speaking about should be considered California specific as that is the only domain of my knowledge

The "dig once" policy is a public subsidy for broadband in the same sense that a national roads network is a subsidy for FedEx or Amazon.com -- it indeed reduces the costs that operators need to pay to get into the market but there's a panoply of benefits to the public beyond those of a literal, "paying government money to broadband operators" subsidy.

Infrastructure, whether broadband or highways, has positive externalities that are hard for infrastructure builders to capture, so the free market under-invests in infrastructure. However, government investment in infrastructure also is distortionary and needs to be considered in every instance.

For example, it's great that we can get products delivered overnight from Amazon. However, as you point out, that is effectively a public subsidy of Amazon's business model, and that has negative effects too. Subsidizing public road infrastructure makes shipping products across the country artificially cheaper compared to making and selling them locally. That essentially strips away an advantage local retailers and manufacturers would otherwise have against big, national companies like Amazon.

> That essentially strips away an advantage local retailers and manufacturers would otherwise have against big, national companies like Amazon.

Indeed -- my hope for "dig once" policies would be that they have an opposite effect on competition; namely, that national-scale incumbents (who can currently afford to throw a million dollars here or there to put in fibre of their own in a new region/market) can get effective competition from local ISPs (who don't have a comparable ability to throw cash around).

In the UK we've had laws requiring common access to telecoms infrastructure since at least the early 2000s, and its had exactly that effect. In almost any home you have a choice of a dozen or so providers, ranging from super cheap basic providers who'll give you a few megabits and a stingy data allowance right up to business specialists who can provide multiple redundant lines on different backbones with 4G backup, generally without any sort of installation being required because. We've also got a big cable provider who run their own lines and give pretty competitive pricing because of all the other options.

In theory Amazon/FedEx pay for their share of road maintenance costs, via gas taxes. In reality, gas taxes apparently only cover about half of such costs (overall) [1], but that still doesn't sound like a huge distortion to me.

[1] http://www.frontiergroup.org/reports/fg/do-roads-pay-themsel...

It's even worse than that, gas usage (tax revenue) scales linear with weight at low speed. Road damage scales as the fourth power of axle weight! Large trucks pay nowhere near the cost they impose on our infrastructure.

I agree entirely. If you're building a road or doing construction, why not put in conduits that allow installation or upgrades for years to come, rather than having to dig repeatedly? And if you have to re-dig an area because it doesn't have conduits, don't just bury a wire, install conduits.

This is directly the job of local government: to build infrastructure. And this wouldn't force the use of any particular service; rather, it would make it possible for many services to exist.

I'm definitely on the same page - that said, a big determining factor of success is the efficiency of the local gov in question. For example, dig once already exists in San Francisco (http://sfgov.org/dt/dig-once). I've had a conversation with someone who project manages fiber installation across the city, and he mentioned it takes many months for any dig to actually happen, because every potential stakeholder has to sign off before going forward.

Again, am on the same page that dig once makes a lot of sense, but there are some practical roadblocks that could make it less desirable.

And you can even create a deployment tax that pays for the tiny cost at a later date when the conduit is actually used.

>> "It creates a slight public subsidy for broadband"

Maybe it's me, but $3 billion is not a small amount of money:

https://news.ycombinator.com/item?id=13941757

$3B is pretty small on the large scale. Especially when you consider the cost of adding them later (while it may not be paid for up front by the government, costs of doing construction, like closures, do add to gov't costs)

According to some sensationalist stories I've read, it's only ~10X what we're spending to protect/transport President Trump on his weekend vacations.

What does this contribute to the conversation?

Someone said it's not a lot of money. Someone else said it was. I provided a relative reference. Not the most monumental - but my intention was to add perspective.

Context. Reasoning about government budget-size numbers is difficult for most people.

In normal human terms, yes, but 3B is only .08% of the federal budget.

Also, there is some economic gain from this activity...paying construction workers to install it, and whatever the end users decide to do with it. Is that gain > 3B? I don't know.

Reply on news.ycombinator.com