Frequently Asked Questions (FAQs)

“Broadband” refers to a high-speed communications network connecting users in homes and businesses to remote resources such as data centers, the Internet, television, and telephone services. We all use one, probably every day, every time we access the Internet or make a call with our smart phone.

The “band” refers to a channel, like a “radio band”; think of it as a kind of digital road.  “Broad” suggests large or fast or substantial in capacity.  The word has been applied to data networks to distinguish high capacity networks from older networks reached through dial-up modems (the word itself is borrowed from old radio systems that divided their frequency spectrum into bands, as we see for example in FM radio).  The FCC defines “broadband” now as 25 mbps downstream (from the network to you), 3 mbps upstream (the other way).  The downstream speed can support three simultaneous HD video channels.  However, speeds are going up, irresistibly.

“Broadband” does not define the kind of communication links inside the network.  A broadband network may be fashioned from legacy telephone lines, legacy cable television lines, new fiber optics lines, satellite links, or mobile network links, the last two using the air as medium.   In our region 65% of broadband networks use legacy cable television lines.  While fiber optics lines exist everywhere in Connecticut for “back-haul” connections from a local network to the Internet, we have no fiber optics system in Connecticut yet for residential broadband connections.  We are sadly unique in that sense.  Fiber broadband is the future.

To begin, a word that feels like it should be plural but is singular.  Fiber optics refers to thin strands of purified glass, the size of a human hair, that when fitted with transceivers on each end can carry modulated light signals as much as 20 miles without amplification.   Its principal virtue is speed, at least 10,000 times the maximum rate one can realize over copper-based wiring.  But fiber optics also emits no electromagnetic radiation, suffers no errors caused by electromagnetic noise, and lasts longer in suitable sheathing compared to copper based lines. It is the wiring of the future.

A modulated carrier requires frequency space around it to create a pipe for the signal after it is modulated.  The size of the pipe is roughly proportional to the rate at which the carrier signal is modulated.  An 88.5 MHz FM carrier requires 15 kHz of bandwidth around it to transmit typical voice frequencies up to 15 kHz.  The physics of copper wire imposes limits on maximum bandwidth.  A relatively short coax line within a cable television network can transmit data at rates no greater than 10 gbps (billion bps), and that only under ideal conditions (noise, power levels, other distortions).  Light ranges from 430 trillion Hz to 750 trillion Hz. If the entire range is used with binary modulation, a medium transmitting light can realize data rates of 320 terabits per second (trillion bps).  No one has figured out how to make a transceiver that would run that fast.  Google announced recently an undersea cable with 18 terabit per second transmission rates, using the highest speed (and very expensive) transceivers on the market.

Fiber optics technology came into being in 1971 when Corning figured out how to manufacture glass with sufficient purity.   Transceiver technology followed, using laser and LED light sources.  Beginning in the early 1990s telephone and Internet companies began installing fiber optics to connect their office and switching nodes together around the world; tens of millions of miles exist now in this “back-haul” system for the Internet and Public Switched Telephone Network.  Much of that wiring is still “dark,” meaning not in use yet, so great is the capacity.  Fiber optics wiring also connects massive computer arrays together in huge data centers.   However, in the United States fiber optics has only recently become a medium of choice for the last mile into homes.  To date carriers have connected about 20 million homes out of some 128 million in the country with fiber optics.  None of these homes are in Connecticut.

“Fiber optics broadband” means a broadband network that uses fiber optics lines to connect premises to the back-haul system we call the Internet.  This intermediate connection is the “last mile.”   It is distinguished from the last mile of broadband networks that use legacy coaxial cable and cable modems, legacy telephone wiring and DSL modems, or satellite links.

A broadband network may be thought to have three interconnected parts: (1) a home network usually comprised of a wireless WiFi system that connects user devices to a WiFi router, with a modem (modulator-demodulator) of some kind that converts digital signals from a router to analog signals suitable to a wire or wireless medium; (2) the “last mile” connection between the home or business and the “back-haul” network; and (3) the back-haul network comprising millions of high speed switches interconnected by fiber optic lines operating at very high speeds.  The “last mile” may be legacy telephone wires with DSL modems, legacy cable television lines with cable modems, fiber optics lines with fiber optics transceivers, or satellite links where the “last mile” is 23,000 miles to and from a geocentric satellite.

The basic difference among the last-mile lines is speed, or data rate.  In our region DSL telephone lines go no faster than 6 mbps and many will be much less.  Cable television lines can get as high as 300 mbps but are shared by as many as 100 users, so will depend greatly on the number of simultaneous users (and hence time of day and month). Satellite links come in below 25 mbps, depend greatly on the weather, and have round trip delays that kill many applications; as they cover anyone and can connect anyone without wiring, they are the link of last resort for those who cannot get broadband any other way. Fiber optics lines have so much raw capacity that the limit on data rate arises from the transceivers that convert electric digital signals to light.  “Last mile fiber” lines can get 10 gbps today, but fiber lines in the back-haul network have rates now as high as 18 terabits per second (trillion bits per second) for undersea fiber optic cables; you don’t want to know how much those transceivers cost.

“bits per second” (bps) means the number of binary digits (bits) per second

“kilobits per second” (kbps) means a thousand bits per second

“megabits per second” (mbps) means a million bits per second

“gigabits per second” (gbps) means a billion bits per second

“terabits per second” (tbps) means a trillion bits per second

Digital information represents things with strings of “bits” or “binary digits” that have only two logical states—1 or 0, on or off.  In most protocols and memory systems bits are fitted together in groups of eight, called a “byte.”  For example, alpha-numeric information is generally shaped according to a standard (called ASCII) that assigns a combination of bits in a byte to letters, numbers, and special characters–upper case M is “01001101”.

But things like video streams will look rather random even when organized in bytes.  A video stream typically (today) starts out as three bytes or 24 bits for every pixel on a screen to express color and intensity and some other things.  An HD screen (1920 x 1080 pixels) requires almost 50 million bits to paint a screen once.  If painted 30 times a second (interlaced scanning) the equivalent data rate is 1,492,000,000 bits per second (1.492 gbps); if painted 60 times a second (progressive scanning), the rate is 2.905 gbps. However, networks are not this fast yet.  All video signals sent to homes are radically compressed, with loss of information in the process.  Standard video requires around 1.5 mbps, HD video around 7 mbps, 4K video around 30 mbps, and 8K video around 90 mbps.  Video today is 75% of Internet traffic, split between standard and HD (most YouTube video is Standard).  These signals are “real time,” meaning they cannot be delayed much in the network, else buffers at various places in the network run out of material, or get overrun.  When Netflix stops streaming, the little wheel of dots whirling around, it usually means network congestion has caused buffers in the network to have nothing or too much.

The other index of value in assessing data rate requirements concerns sizes of files being sent.  A single compressed HD two-hour video requires about 6.3 billion bytes.  Such a file requires about 2.4 days to transmit at dial-up speeds of 30 kbps, about half an hour at 25 mbps, and about 14 seconds at 1 gbps.  If you have filled up a one-terabyte file on your computer and want to send it to someone, or a data center for back-up, it will take 3.7 days at 25 mbps, 5.5 hours at 1 gbps, and about 33 minutes at 10 gbps.  These size files will be commonplace in not too many years.  Fiber optics will allow you to be prepared; no other network will.

Here’s a word thrown around the data network world you may not have heard before. But it is becoming important.  “Latency” refers to “round trip delay,” the time it takes a bit or packet of information to go from your device to a remote resource and back.  It is sometimes called a “ping,” usually on speed tests systems that measure upstream and downstream data rates and round trip delay.  Systems measure latency in milliseconds (thousandths of a second).  Typical latencies in land line and mobile networks today range from 15 to 75 ms; they are traffic dependent.  As it takes a signal 125 ms to reach a geocentric satellite, best case, latency in satellite connections is around 500 ms, or half a second, best case.

Except for satellite connections, latency has not been a critical performance parameter in common data networks.  One-way video streams hardly care, file downloads and uploads are not critically affected, and even web page building, a process with calls and responses, does not suffer adversely with today’s common latencies. So why are 5G zealots waxing lyrical about 1 ms latencies from future 5G networks, and why should wire-line networks care as well?  The most common answer concerns new marque applications—autonomous vehicles, remote Virtual Reality, remote surgeries, and future telemedicine machines.  When and if 5G networks support autonomous vehicles, delays between data from the car and control signals back to the car can be life or death.  Not only must the delays caused by the network be much lower than delays from today’s network, they cannot be traffic dependent.  Virtual Reality screen rebuilds have to happen within 10 ms of a head moving to prevent nausea; this must include both network delays and processing delays, so the network delays have to be really small.  Remote microsurgeries are in the same boat; requiring rapid and reliable responses to surprising information from inside the patient.  And telemedicine machines of the future will need to transfer huge files of data without error very quickly; such files must be transmitted in small groups with error control on each group, making round trip delay a critical part of the performance equation.

How to achieve such latencies is far from trivial.  Current 5G trials have not reported anything better than 7 ms.  Any connection that goes through a large number of switches will fail because of processing delays at the switches, and at 1 ms for every 186 miles traveled the delay in signal transit becomes material; therefore all such applications will require massive numbers of edge data centers within a hundred miles or so of any user instance that do not exist yet.  And how to make such delays traffic independent without network overbuilding that cripples any sensible capital budget of network builders looks like a problem for Solomon.

Depends. (1) If you do not have a cable television connection today (in our region), you will need a next generation network to watch Netflix or any other form of HD video streaming, plus a host of other things that slow speeds make very annoying or impossible.  (2) If you have cable but suffer slowdowns with video streaming or video games or file uploading or anything else that annoys you, a next generation network operating at 40 times the speed downstream and maybe as much as 120 times the speed upstream will warm your heart.  (3) If you are happy with your cable television capacity but have bad memories of Comcast, Cablevision (Spectrum), or Charter (Optimum) you will be glad to know that a community network will have community service, a real person who likely lives in the region.  (4) If you are happy in general with your CATV network and service, we can only say that a new network will be next generation and the last network you will ever need, letting you feel good about your future.

However, even if you are happy with your network today, there is very likely a time in the future when you will not be.  Cable television companies must upgrade their networks, at considerable expense, to keep up with increased data rates.  This usually means extending the frequency range of each line to a group of customers (as many as 100 share every line).  Extending frequencies requires new amplifiers and very often shortening or reconditioning lines themselves, the latter a significant engineering and field operation.  Our region’s networks have not been upgraded for a long time, even compared to other locations in Connecticut.  You can get a nominal peak service in Hartford or San Francisco at 1 gbps; tops here is 300 mbps, for just a few people on any given line.  A new fiber optics network will start out at 1 gbps, in both directions, and upgrades only require new transceivers at each end—no in-network modifications.

Let us count the ways.  Restoring our young population.  Economic development generally. Preparing for future in-home health care systems.  Repairing the digital divide in education.  Creating a network that optimized digital education for all.  Bringing farms into high speed world that is becoming increasing necessary to reach markets and acquire information.  Improving homes values and sales.  Showing Connecticut how it can be done (the state needs it badly).  In general, fashioning for our region something that has, in a very short span of time, become as necessary as roads and electricity—broadband connectivity for all.

Of these reasons the first must stand out.  Our population of people aged 20 to 39 is one-half what it should be, and declining, in absolute and percentage terms.  It will take in the order of 8000 new jobs to correct the imbalance, a 25% increase over regional jobs today.  Furthermore, they should be well-paying jobs.  The only way we are going to fill such a need is by attracting high tech companies and jobs in other companies here.  The promise that every home in a town has gigabit service in both directions—available nowhere else in Connecticut now and in very few places in America, including the San Francisco Bay Area (not a single city)—will operate like a magnet.  We have to do other things; we are doing other things.  But the network will be the lynch pin, the symbol of connecting our communities to the future without losing its beauty, recreation features, and elevated quality of life, not to mention affordable housing compared to New York or Boston.

This is not a silly question, but it may be the wrong question.  It seems automatic—if broadband has become as necessary as roads and electricity, and it has, then everyone should have access to it.  But roads and electricity are already ubiquitous; broadband is not.  35% of our region’s homes do not have broadband as understood by the FCC.  Some limp along with DSL lines, some with satellite connections, some with nothing at all beyond terrible cell phone service.  We are rather like where we were with electricity 120 years ago, some having it, some not.  The question then, as now, is not so much why, but how.

We actually make a stronger claim: everyone should have access to the latest technology.  Electricity has always been 120 volts at 60 Hz; roads have always averaged 12 feet in width.  But data rates grow with every circuit of the sun.  What was good ten years ago is woefully inadequate today, and what seems adequate today will seem terrible in ten years.  The only network capable of tracking future demand is one built with fiber optics to the premises.  Even cable tv companies know this (the reason CableVision is threatening to upgrade its American network to fiber optics by 2022).

There will be initiatives in our region to connect parts of towns with fiber optics, leaving the least dense areas for another day.  We do not oppose such efforts; indeed, we encourage anything moving in the right direction.  But the power of universal service should not be ignored for those who do not live here now, but will want to.  We suffer a digital divide now, some homes so deprived of network access that they cannot be sold.  Our prospects for new young people go up, perhaps significantly, if we can say, honestly, that any home has or will have very soon a gigabit connection with fiber optics.  To do less is to transfer one digital divide for another.

The simple answer is that communities will not get a universal next generation broadband network without subsidies of some sort.  Unfortunately in our region all but one form of subsidy are ruled out.  That form is local borrowing paid back through taxes.  It will be just like roads and schools and airports, something the community must support in general if the service is accessible to everyone, as it should be.

The hard truth is that a fiber optics network that connects everyone will require subsidies of some sort no matter where  it is constructed.  It has ever been such with telecommunications to homes as long as we insist that every home pays the same amount (a condition of “affordable”) when the costs to connect homes vary tremendously.  The major costs come with wiring. Homes bunched together near a switching node are much less expensive to serve than homes miles away and spread out; and wiring underground in a major urban area is way more expensive than putting wire on poles.   Subsidies come in various forms: federal grants; state grants; cross-subsidies within an electric utility that also operates a fiber optics data network; overcharging business clients to hold residential customers within affordable range; free access to municipal buildings and infrastructures along with tax breaks; local borrowing repaid through taxes.  But subsidies there must be.

We are not eligible for federal grants because we are too well served by cable television networks.  The state has no money, and states with money have tended to adopt the same requirement for grants—no service at all.  We do not have any electric utility in our region. We do not have enough businesses here we could plausibly overcharge to pay for residential service.  And we do not have enough existing facilities a private network operator could use that would offset enough construction and operating costs to relieve our communities from direct subsidies.  It is a painful choice, but we either limit service to those in more concentrated areas of our towns, or we must summon money through some form of taxation to enable service to everyone.

For the network, almost certainly not.  But there may be other ways to get federal money into our region.

The federal government has two major programs devoted to rural broadband.  The Connect America fund, created through a percentage tax on every phone bill, spends about $50 million per year through FCC grants.  The Reconnect America fund, administered by the Department of Agriculture, funnels federal tax dollars to rural broadband networks to the tune of $200 million a year now, with another $400 million in loans or loan/grant combinations.  The preponderance of this money today goes to rural electric utilities, small rural telephone companies, or cooperatives of either.  The condition is the same for all—no existing broadband service.  The money only goes for construction; the networks must be self-sustaining thereafter, one major reason for money going to existing utilities.  With cable television connecting 65% of our homes, we will never be sufficiently “unserved” to qualify.

However, installing a fiber optic network may have collateral benefits in this respect.  There are numberless resources now for subsidies involving health care, the digital divide, other educational advances for those from the economic margins, agriculture, developing government efficiencies, aging at home, and combatting poverty, any of which will be promoted in our region by a fiber-optic nerve center linking everything together.  The existence of the network will make us a more attractive opportunity for these related concerns.

Municipalities pay for and own the trunk wiring on poles (think roads) and a private partner, or partners, provides the wiring from the poles into homes, all network electronics (think driveways and garages), operations and maintenance after construction, and services.

Our small-town populations and relatively small number of homes spread far apart from each other prevent any sensible business case for a private partner if the partner builds the entire network.  But if we install and own the wiring on the poles, the largest cost in our region for a new network, and the private partner provides the rest—drop wiring from the pole to the home or business, home electronics and wiring, and network switches—the business case for the private partner becomes attractive—after some initial infrastructure he only commits capital when he has a customer.  This model greatly eases the risk burden for the private partner and creates the conditions for good working relationships between the public and private parties.  The model is unique in America as far as we know despite it making a great deal of sense, functioning exactly like the way we treat roads and sewer systems—trunks owned by the municipality, home facilities provided by private parties, including home owners.

Several.  The key opening question is about universal service—yes or no.  The only reason to answer this question in the negative is to avoid subsidies to pay for construction, operations, and/or  maintenance.  The most common model is the American tradition—private carriers such as Verizon that requite all costs through revenues and enjoy local monopolies.  The second most common model involves a local utility—electric or telephone—that obtains subsidies and requites all operating costs through revenues.  A third model involves public/private partnerships but with revenues alone paying the bills; such networks will seldom if ever be universal.

It costs in the order of $1500 per home to construct a new fiber optics network in a large city using telephone poles.  It can run as high as $5000 per home if underground.  It can cost as much as $6000 per home in very rural areas; it will cost in the order of $4500 in the more rural communities in our region assuming everyone gets service.  But the costs per home can be as low as $2000 in select areas of rural communities with something like a town center.  To sort out the right business model, a rural community faces two central questions: (1) do we serve everyone, and face the inevitable duty of public funding; and (2) how do we provide operations and maintenance (O&M).  A rule of thumb suggests that private carriers—incumbents or new—can make money if per home costs come in at or below $1500, and there are enough customers to defray O&M costs through revenues.  Answering the second question depends greatly on whether O&M is all new or can leverage operations already in existence for other things.   It should not be surprising that existing fiber optics networks in American come from incumbent carriers in large cities or networks operated by an existing electric or telephone utility that has revenues from other sources.

We have four possibilities for our region, one without any existing utilities but saturated with cable television broadband and hence barred from federal or state subsidies.  (1) One is the construction from scratch of a regional utility that serves enough customers to repay all costs through revenues.  The hurdles for such an idea are so huge as to forbid any further thought even though the numbers might run.  (2) Another is to entice a private party to construct and operate a network for sections of towns with premises close enough together to make business sense without public subsidies.  (3) The third is the model suggested herein, of a public/private partnership in which communities defray some costs through local taxes.  (4) The fourth, a variant of the third, will be odd but worth a little consideration.  CableVision, owned by a French company (Altice) has announced intentions to upgrade its networks around the country, serving 8 million customers. Cablevision provides networks and services for eight towns in our region, including Litchfield and Torrington.   Even if they decide to upgrade here (odds are against), they will not connect everyone in their own territory nor are they likely to invade adjacent territories.  However, they might be induced to connect everyone, and invade adjacent territories, if communities supplied the trunk wiring (the only reason they do not connect everyone).  Think about it.  Could be really cool.

This is the key question.  Everyone wants the answer to be zero.   But if everyone wants to be connected, and communities decide everyone should be connected, like we are connected to roads and electricity, then some moneys will have to flow from every household to build the network.  Exactly how much depends upon too many factors to present in this section, but if the entire region does it together the costs could be less than $10 per month per home.  If one rural town does it alone, the costs could get as high as $20 per month.

Network costs cannot be determined or pinned down until networks are fully engineered, and that only happens after a contract has been signed.  So we make educated guesses.   We are assuming for this purposes that communities adopt the proposed business model in which communities own the trunk wiring on the poles and a private partner provides all drop wiring (from pole to home) and all electronics.  We are assuming that wire alone, with no electronics, will enable communities to obtain 40 year terms for loans.  We assume $50,000 per mile for installing fiber optics wiring on poles.  This also assumes several variables that could move the number in both directions, but not by more than $15,000 per mile.  If the entire region did it at once, with a network passing 76,000 homes over 2200 miles of road at $50,000 per mile the cost per month per home would be $6.05.  If Torrington did it alone, with 16,600 homes and  204 miles of road, the costs per home would be  $2.72 per month.  If a more rural community such as Norfolk did it alone, and wired poles passing every home, the costs would be $17.20 per month.  (These are simple calculations pushed through a business calculator and do not take account of many contingencies, including the prospect that some leasing arrangement could be made with the private operator for a regional or even a local network in Torrington, where the housing densities justify private costs for O&M.)

These are costs borne by every home and business in the community.  It could be in the form of an increase to the mill rate and paid yearly through property tax bills.  It could be assessed through a separate municipal entity much as sewer systems are assessed.  The latter would enable the costs to be pro rata, the same for every home, rather than progressive, in proportion to the assessed value of your home.

Internet access at gigabit rates and VOIP telephone services.  The system in the beginning will not offer broadcast television services.  Over time we imagine some auxiliary services such as local data centers for applications requiring very low latencies and private, secure connections to private data centers associated with corporations supporting work at home.  The latter two would enable so-called zero-clients, a desktop with keyboard and monitor but all computing done remotely—a workstation for $200.

The base service of course will be Internet access at symmetric gigabit rates—one billion bits per second in both directions.  This will support the vast majority of needs for a decade or more. Upgrades to 10 gbps will only require new transceivers at each end—a simple exchange of boxes or cards—without touching the in-network wiring.   The base service will also include landline telephone connected to existing home telephone wiring.

As Work at Home for individuals working for large corporations represents one of the most important targets for a new network, we will explore means of connecting home networks to distant corporate data centers through secure VPN connections that would enable users from home to have the same data rates and latencies as they see at the office.  There will also be significant opportunities for zero-client desktops, ones comprising a keyboard and monitor but no computer; with gigabit speeds and low latency to a local data center, zero client desktops can realize work station capabilities ($5000 computers) at the cost of the keyboard and monitor.  These are common now in corporate work (users cannot mess up their computers, the largest maintenance cost for IT departments), and would be a great favorite for students and individuals on the economic margins.

Sadly, we cannot say now with any precision.  Services will be provided by our partner or partners; they must make money.  Service revenues will be their only reliable means.  However, we are targeting costs for Internet Access within range of current costs of $60 per month for 1 gigabit rates.  Telephone services will be under $20 per month.  Business services of course will be considerably more (businesses use considerably more network capacity) but in our region they are unlikely to fill a gap.

We are a small market with many miles of road and relatively few homes.  Even if we pay for the trunk wiring, a partner must spend $1000 or more to connect each subscriber’s home, and then he must have trucks, buildings, inventory, and people to maintain the network, change out or move customers, bill customers, answer service calls, and provide a level of administration that makes the operation hum.  He needs some minimum number of customers to make a business case.  It is clear that a regional network passing 67,000 homes would be more than enough for the numbers to run and service costs to come in around $60 per home per month for Internet access.  How this might work out for one town the size of Norfolk is still in discussion.  Clearly the promise of several communities going ahead with network improve the picture.

Of the estimated 20 million homes in America (out of 128 million possibilities)  with fiber optics connections, more than half have been wired by an incumbent carrier—Verizon, AT&T, CenturyLink, or Frontier—all in large urban areas.  On the other end of the scale, some 150 of the 1250 electric or telephone cooperatives in America have installed fiber optics networks with funds supplied by the federal government, with another 100 likely to be funded per year over the next decade.  A number of midsized municipalities with existing electric utilities have constructed universal fiber optic networks, led by the poster child, Chattanooga, all subsidized in one way or another through the utility.  Some number of rural towns, less then 200 in total, have cobbled together networks with state or federal subsidizes, using contractors for construction and O&M; several towns in western Massachusetts number among the fold.

It is fair to ask in this context why so few have been wired when a few other countries have already wired almost everyone, and some, such as China, are moving swiftly with federal programs dedicated to universal connectivity.  The answer lies in two independent areas of our national condition.  One is that we have broadband over cable television networks available to more than 90% of the country now, a level of penetration unique in the world.  More than 50% of homes in America subscribe to broadband over cable lines at rates at least of 25 mbps, and many are well above 100 mbps (a typical data rate in other countries for fiber connections).   In urban areas the two giants in that business—Comcast and Charter (80% of market)—are offering 1 gigabit peak services, way more than is needed for most applications today.  Another significant number obtain true broadband (25 mbps or above) through existing telephone lines and advanced forms of DSL modems.  While an exact analysis is beyond reach for this report, we can also say with confidence that the American business and academic establishment has more fiber optic connectivity than any country on earth, by far.  We are the digital capital of the world, still.

The second reason goes back to 1215 in England, when nobles forced King John to account for his use of tax money.  We are very averse to taxation.  There are many good reasons to be averse to taxation, but at times it gets in the way of community well-being.  This is such a time and issue.  We need advanced digital highways to every home and business in America, sooner rather than later, for all kinds of reasons that go beyond our station in the world.  And, unlike roads and sewer systems, which are extremely expensive, the community costs are very low, the price of a cheap meal per month.  This is not to deny some problems with support for those who really cannot afford it, or a sense of equity across widely divergent incomes per house.  But the price is very low for the gain, which over time will be very high.

No.

First, 5G will not materialize in our region for decades without serious subsidies; it is in very small trials now and the standards for some of its features have not even been set yet. Verizon has stated explicitly that they do not plan to even bring trunk wiring to rural parts of Connecticut.

Second, 5G requires small antennas on utility poles or new poles spaced as close together as 300 feet to achieve its top priority capacities.  Installing these antennas everywhere will be expensive, time consuming, and opposed by many communities.

Third, connecting all these antennas back to Core Networks requires massive fiber optic cables, with 8 strands devoted to each antenna, a time consuming and very expensive task that is barely started in a few urban areas.  Wiring that will take three more years has begun in Hartford. Count on eight to ten years to wire out the I95 and I91 corridors, if then.  (It was recently announced that AT&T is cutting $3 billion from its capital budget for each of the next three years, and Verizon has a smaller but material roll back as well.  Reality is setting in.)

Four, none of the marquee applications that have prospects for generating the incremental revenues necessary to justify the tens of billions of dollars in capital expense exist yet, even in trial form—autonomous vehicles, remote Virtual Reality systems, remote microsurgeries, high capacity telemedicine machines, the Internet of Things with 50,000 things in range of every antenna.

Five, even if 5G were everywhere, it would be a poor alternative to bringing fiber optics all the way to the home.  Each 5G antenna will be shared by tens if not hundreds of housing units and up to 50,000 things in the Internet of Things (based on a target of 1 million per square kilometer).  The same fiber wire brought a little further, into the home, will only be shared by the devices in that home, without all the expense of the antennas, at data rates that are achieved in homes today of 10 gbps.  5G is not even close to that figure today, and will not be able to deliver that rate to many subscribers at the same time perhaps forever.

We will have 5G and Fiber to the Home everywhere someday.  Fiber is possible today; 5G is a distant dream.  We should bet on today.

Twenty-five percent of roads and homes in our region do not have cell phone coverage.  It is easy to say that this is unacceptable (it is), much harder to say what we can do about it.  The carriers own the spectrum used for wireless signal transmission, and require their own equipment  to connect antennas to Core Networks even if communities owned the antennas.  They have shown no interest in doing this (no incremental revenue to justify the incremental costs).  However, we are thinking about a 5G play that might bring them to the table, one in which we supply the fiber optics (the most expensive part, which we will use for our Fiber to the Home networks as well) and they supply the antennas.  Could be a win-win as they say.

Cell phone  companies know they cannot install 5G in rural regions of the country.  They are all but begging Washington for subsidies, with no success so far.  The problem is antenna density versus revenue.  To realize its stated long -term objective of 10 gbps peak data rates, 1 ms latency, and 50,000 devices per antenna, 5G cell phone networks must radically densify, by some estimates placing antennas 300 feet apart (or even closer if some early trial reports are correct).   While AT&T is presently crowing about 5G with antennas having ranges around 2 miles, they hide the fact that these antennas and frequencies are not 5G, really, but a kind of grown-up 4G, data rates in the low hundreds of megabits per second, but miles from the top of their mountain.

While the antennas themselves are expensive, the major cost for deployment is the fiber optics wiring required to connect each antenna back to a Core Network.  Suppose we offered to provide the wiring?  If we used the same wiring that we were installing for our own broadband networks with fiber to the home, the cost to us is small—many more strands in the cable, but that will add no more than 10% to our costs, a trivial amount if universal mobile coverage was the outcome.   Our region must have more than 100,000 mobile subscribers.  That must be interesting to someone.

If you hug an operating antenna for long periods of time, perhaps.  Otherwise, they are less harmful, if they are harmful, than the signals produced by your smartphone.  Indeed, the closer you are to a mobile antenna when using your smartphone, the lower will be the amount of radiation absorbed by your head and body.   Unless you are willing to give up your smartphone, you should be lobbying hard for small cell antennas on telephone poles to the extent necessary to provide universal 4G or 5G coverage in our region.

The science on low-frequency radiation is far from consistent or complete.  Many studies suggest harm at certain levels over sustained periods of exposure; many others suggest otherwise.   All but one of the epidemiological studies show no effect from the 700 billion smart phones and their antennas sold over the last decade.  As these studies usually involve high levels of self-reporting, they are all suspect, including the one from Sweden that does show an increase in cancers that could have been caused by radiation over the last twenty years.  These studies are all correlations at best anyway; establishing causes is much harder.   We are never going to test human beings in controlled studies (to find out what levels cause cancer and other maladies), nor can we ascertain now what effects radiation will have on long term prospects for cancers that may take decades to materialize.  Perhaps the best clue is that no government agency in the world has declared that the FCC standards for maximum transmit power levels from smart phones and antennas are too high.

Energy levels from mobile antennas drop rapidly as waves move from the antenna to a user.  The level thirty feet away is 1% of the original.  Mobile phones expect receive levels in a range from -40 dbm to -115 dbm (the former is 0.0000001 watt, or eight orders of magnitude below the 40 dbm (10 watts)  transmitted by a small cell antenna.  Meanwhile signal levels transmitted by a smart phone range from +30 dbm to below +10 dbm—1 watt to 0.01 watt.  The key dependency is receive power; the higher the received signal (the closer the antenna), the lower the transmit power, to save battery life.  So, if you want to continue to use your smartphone, you are probably in no real danger, but you will be better off with many small antennas close by than one antenna 20 miles away.

We have already had a form of protest, Frontier and the cable companies persuading our Public Utilities Regulatory Agency (PURA)  to bar municipal access to telephone poles.  We sued, and won a key victory in December of 2019.  Frontier is presently in serious trouble and may not notice until they are reformed into some other company or shape.  The typical response of cable television (CATV) companies to a municipal network has been to offer discounts on longer term contracts.  However, the experience of other communities with municipal networks and competition from cable companies is encouraging, with 50% or more of homes taking services from the municipal network rather than incumbent carriers.

Incumbent carriers spend considerable sums to interfere with competition, in the political arena, in the courts, and in the marketplace. They have been successful with state legislation curbing municipal networks in many southern states.  But Connecticut has a law, number 16-233, that entitles municipalities to use existing utility poles “for any purpose” without paying yearly pole fees, a decent bargain since, unlike municipalities in other states, Connecticut municipalities do not charge pole owners for use of public rights of way.  Frontier and cable companies talked PURA into restricting “for any purpose” to municipal purposes.  Three entities representing municipalities sued (including Northwest Connect), and won.  We are free to build municipal networks using existing utility poles without yearly fees.

What happens next can only be conjectured.  Incumbents may have lost their appetite for litigation.  The current composition of the Connecticut legislature makes a legal recourse unlikely.  Frontier has almost no market power in our region.  However, the three cable companies in our region can be expected to react when we start making network announcements; they enjoy their monopoly power here.  The most likely offense will be to offer 20% lower rates for two year contracts, what they have done elsewhere.  This will do nothing, of course, to the  35% of our region not served by a cable tv company.  But experience in other towns with municipal fiber optic networks suggests that the effort does not prevent a municipal network from getting 50% or more of homes to subscribe to their services.  Much of the reason goes to customer service.  Cable tv service is terrible and unlikely to get better.  A local call center with real human beings who live in the area answering the phone without a robot answering first feels like a tonic after torture.  In short, this is not a huge worry.