Current Affairs

A Series of Tubes: Reclaiming the Physical Internet

Who owns the internet? Everyone but us. Sparky Abraham makes a case for nationalizing the infrastructure of an increasingly online world.

If you’re like me, you are largely oblivious to the world around you. The knowledge of the entire world, along with a direct line to everyone you know and love, lives in your phone or your computer and you carry those around with you. Occasionally you plug them into some magical holes in your wall so they keep running. And of course you need wifi or cell service, which floats around in the air as long as you’re in the right sort of place with the right password. If you can grab it, then there you are. Plugged in. Connected, networked, internetworked, webbed to the wide world. The internet is in the air. It’s a cloud. We’re all connected, somehow, by technology or something. The magic lives in our computers and phones and in the apps they run.

There’s this anecdote in Michael Lewis’s Flash Boys: After most of its trading goes online, a hedge fund decides it might as well move from Manhattan to the midwest. Physical location has ceased to matter because business lives in cyberspace now. What a boon! What a time to save on rent. Of course, this turns out to be exactly wrong. As this firm is trying to save money by moving west, the high frequency traders on whom the book is focused are spending hundreds of millions of dollars to get closer to the market, to put their computers as close as possible to the market computers. 

If you’re a hedge fund and you think about the internet like I usually do, you move to Oklahoma where the rent is cheap. If you think about the internet the way it actually is, you pay huge sums of money to put your computers inside New Jersey warehouses directly next to the market servers, and even more money to get them a few feet closer than your competitors’ computers. If you think about the internet the way it actually is, you spend untold millions to lay a few strands of glass, no thicker than human hairs, in the straightest possible line from Chicago to New Jersey. If you think about the internet the way it actually is, you worry about mountains and rivers and oceans, about electricity generation and signal interference. And you might also worry about ownership, control, power, and coordination. 

The internet, after all, is simply a communications network. Its job is to get information from one physical place to another physical place. It serves the same function as telephones, telegraphs, radios, smoke signals, and semaphore. It happens to operate through a complicated system of physical tubes. When you send a packet of information in the form of a letter, a federal government agency takes your letter and carries it across an intricate nationwide network of federally-, state-, and locally-owned and managed roads. When you send a packet of information in the form of bits, it is usually carried by a privately-owned internet service provider (or “ISP,” usually the local cable monopoly) along an intricate nationwide network of pathways owned by that company and a myriad of other private companies. It’s like if you could only send a letter with either UPS or FedEx (depending on where you live), and the letter had to travel on nothing but toll roads owned by a few dozen private corporations to get to anywhere from down the street to across the continent.

This presents both a problem and an opportunity. The problem is that as more of our lives move online (whether in the normal course or all at once thanks to a global pandemic), we become ever more at the mercy of the private entities that control the internet. In normal times we have to use roads to get to work and to school, but at least we own the roads and exercise some control over them in the form of local government. We don’t own the internet. There are no city council meetings where we can go to complain about potholes and service interruptions, or about local tax raises and data rates. The way that our public institutions tend to get more people connected to the internet is to give private internet companies grants to build more wires. But those companies still own the wires, and at the end of the day they tend to do only what is most profitable.

The opportunity is to flip this dynamic. But we must go deeper than the calls for municipal internet service providers that were in heavy rotation around 2017. Municipal ISPs are good, but they are only local. Every city in the country could have its own locally owned and democratically controlled ISP and the internet would still be at the mercy of private profit. Someone still has to connect all those cities to each other. Someone still has to carry the bits from Los Angeles through Las Vegas, Salt Lake City, Denver, and Omaha to the servers in Council Bluffs. Owning local tubes is great, but what we need is public national internet infrastructure. To stay with the road analogy, owning local roads only gets you so far—we need democratic ownership and control over the highways and interstates, too. To see why, let’s talk a little more about what the internet is and where it came from.

The physical internet is a lot more like a road system than you might think. Information superhighway cliches aside, a functional internet requires physical pathways reaching every home and business in the country (and, increasingly, in the world). Like road systems, the physical internet is very expensive to build but confers massive benefits once it is built. Like roads, the physical pathways of the internet are neither very technologically complicated nor likely to change much over time. The ways of packaging and transmitting information are always improving, just like the ways of packaging and transmitting things along roads, but the glass wires that internet information travels through have been in place for decades and will not become overloaded for many more decades. 


When the telephone system was first being built, its challenges were far less technological than logistical. The technology was relatively simple: every household with a telephone had one pair of continuous copper wires connecting it to a local switching center. The phone itself was two devices—one near the mouth that converted the sound waves of a person’s voice to a corresponding electrical signal, and the other near the ear that converted an electrical signal back to a sound wave. Picking up the phone closed that wire’s circuit, i.e.  connected the wires together, creating an unbroken loop back to the switching center, where an operator was on the other end of the line. You told the operator which house you wanted to talk to, and they plugged your two wires into the two wires going from the switching center to that house, creating a continuous loop between you and the person you were calling. 

The logistics of this are mind-boggling. Every single home with a telephone had its own unique pair of wires. Those wires had to make it safely from the home to the switching center. Hundreds or thousands of wire pairs had to be color-coded, tagged, and bundled in such a way that they could be discerned, organized, and repaired at any point along miles and miles of physical distance, strung from pole to pole or running underground. The organization and maintenance of those wires was a monumental task for the telecom industry. And its most significant innovations came not in the form of new types of phones (though there were those), but in ways to cram together multiple conversations into the same sets of wires and to automate the task of the operators at the switching centers.

The internet feels more distant from this past than it actually is. To exist online is to send and receive information. To be sent and received, that information must take some physical form. It could be a patterned flow of electrons, a patterned electromagnetic wave, or a patterned series of flashes from one or multiple lasers. In any case, it has to exist somewhere. It has to be sent and then received, and it has to traverse physical distance between where it is sent and where it is received. For the vast majority of information and the vast majority of the distances it travels, it lives as signals in long strands of metal or glass. Unlike the early telephone system, many households and devices can share space on the same strands of metal or glass, but that metal and glass is the internet. 

The glass—optical fiber line—is a thing of wonder. Fiber has been in widespread use for decades and we haven’t come close to reaching its bandwidth potential. The thing about fiber is that it is simply a medium for signalling with light. And, because the signals themselves are merely light flashes, and because light travels so fast (though somewhat slower in glass than in space), the number of signals that are theoretically possible is very, very large. The only constraint on how much data can travel through fiber right now is how fast we can make a laser flash. As we continue to develop faster sending and receiving equipment, the data capacity of the very same fiber that has been in use for ten plus years will continue to expand.

Art by Ben Clarkson

The internet is not a metaphorical world wide web of connection, it is a physical world wide web of connection. If I want to video chat with Current Affairs editor Eli Massey using a service like Zoom while he’s in Cairo, my computer has to send waves to my router, which converts those waves to patterns of electrons in my home coaxial cable, which heads up to the utility pole behind my building and along utility poles or in trenches to the routers at my local Comcast switching center—sometimes the same switching center the old phone company used, but with the windows bricked in and full of servers and routers instead of human operators—and from there likely along fiber-optic strands of glass to a major regional switching center, maybe in downtown Los Angeles or Las Vegas, and onward still until it hits a switching center where Zoom has some co-located servers (maybe down the street, maybe across the country). Then Zoom sends that data on through more switching centers into an undersea cable (yes! my face converted to 0s and 1s encoded into laser flashes whizzing by beneath the sharks and flounders and whales and jellyfish), up under an Egyptian beach and through some Egyptian switching centers until it hits Eli’s router and gets beamed to his computer. Meanwhile his face is making the reverse journey.


Who is involved in this world wide web of strands of metal and glass carrying bits of ourselves over mountains and under oceans? Your ISP provides connection to your router, so they are the first custodians of everything you do online. But once your signal has passed through your router and into a cable that leaves your home, others are suddenly involved. That cable might be owned by your ISP, or it might be owned by a telephone company, or by your city or state. Your ISP might just be a tenant, leasing out some space (i.e. bandwidth) on someone else’s physical network. Whoever owns the cable, they probably do not own the poles it is strung along—those are often owned by the local telephone company or the local power company, and the owner of your cable probably has to lease space on the pole for your cable. (If your cable goes underground, the cable owner might be leasing space in an underground pipeline with a lot of cables, or the cable might simply lie in a tiny, permit-required trench in the dirt in city-, county-, or state-owned land.) Your ISP might own the whole switching center your data hits first, or it might rent space for its physical routers in someone else’s switching center, or it might rent capacity on someone else’s router in yet another party’s switching center. Then onto the next switching center, on yet more cables, with yet different owners and lessees using yet different poles and/or trenches.

Even if your ISP owns the physical cable in your area and is in full control of the signal and medium from your house to its switching center and beyond, it will have to hand your signal off to some other company at some point. The largest of the glass and metal strand owners are called Tier 1 networks. Roughly, this means they have enough physical reach to leverage free transfer agreements with the other Tier 1 networks. That is, if signals traveling on AT&T fiber need to pass through some CenturyLink fiber to get where they’re going, CenturyLink does not charge AT&T, and vice versa. They both own so much fiber (750,000 miles for CenturyLink, 410,000 for AT&T) that they consider each other “peers” and freely grant access to each others’ networks. Comcast, on the other hand, although the ISP for about 40% of all US internet subscriptions as of 2011, is not a Tier 1 network. It might have free peering agreements with most fiber owners, but there are still places on the internet (remember, physical places) where it can’t send data without paying someone to use their fiber. 

This setup presents a serious challenge for the idea of municipal internet promoted by many progressives. The desire to focus on local control, local infrastructure, and local accountability makes perfect sense when you think of your ISP as connecting you to a magical and disembodied internet. But the fact of the matter is that your city cannot provide you with the internet. It will need to link up with someone else at some point. When you try to access a website whose data live on a server in Council Bluffs, your signal needs to get to Iowa and back. Unless you live in Council Bluffs, your city network is not going to get you there.

So, without some larger democratic structure in place, municipal networks are somewhat at the mercy of the existing internet giants. Even if your local government manages to install a fiber optic connection in every home in town, it will still have to pay at some point down the line to access others’ networks and switching centers. It may have to pay its competitors like Comcast for access to those networks and switching centers, and Comcast will have little incentive to offer access on any reasonable terms to a government entity trying to cut into its profits. This is similar to the classic net neutrality problem, which is as follows: since Comcast is a private company with theoretically no obligations to the public, why can’t Comcast charge people extra to access high-demand websites like Netflix? What if Comcast decided it doesn’t like Current Affairs, and simply stopped providing access to our website at all? And since the companies that own the wires have largely acquired the companies that produce the content that travels along the wires (Comcast owns NBC, AT&T owns Time Warner, etc.), what if internet providers started making their own content much easier to access than the content of their competitors? You can see how this net neutrality problem could translate to physical networks: what’s to stop Comcast (or CenturyLink or AT&T) from charging entire cities more to send signals on the nation- and world-wide networks of wires, or not providing access at all to cities with the temerity to shoulder in on private ISPs’ market share? Own the wires, own the internet.

There are a few different ways we might confront this problem. One would be simply to build hundreds of thousands of miles of public cable and fiber, creating a large enough network to replace the current major players, or at least to achieve Tier 1 status and leverage a free data exchange. No public project has attempted this yet, but this has been the go-to move for tech companies seeking complete internet dominance and control. Google Fiber is an ongoing attempt to build a nationwide fiber optic network, to literally string Google-owned fiber alongside CenturyLink-owned fiber across the continent, thereby capturing a large chunk of the physical space internet for Google alongside its increasing control over the cyber space of the internet. If Google ever succeeds with this project, traffic via Gmail or Youtube could travel from Google’s servers to your computer without ever passing out of Google’s control.

Google isn’t alone in this ambition. Many of the big tech companies that dominate the cyber space of the internet have been trying to find ways to take over its physical space as well. Beyond the prospect of domestic control and profit, expanding internet access to offline parts of the world carries the potential for untold profits. Google had a short-lived project called Loon that attempted to float wireless routers on massive balloons in the high atmosphere to enable connection in remote areas without the infrastructure for physical cables. Facebook attempted a similar project using a giant solar-powered plane, essentially a flying cell tower. Eventually these two projects learned the same lessons as Motorola’s Iridium satellite network and AT&T’s microwave relay tower project: it’s very hard to transmit signals quickly, cheaply, and reliably through the air. Long strands of glass are the way to go.

That’s not to say the air isn’t still in play. Both Tesla and Amazon are in the process of attempting to build networks of hundreds or thousands of low-orbit satellites to replace the strands of metal and glass on the Earth’s surface. The traditional problem with satellite internet is that you have to be able to send and receive signals from a satellite, meaning it has to be somewhere in the sky above you, and not on the other side of the earth. There are only two ways to make sure there is always a satellite above you. First, you could put the satellite in geostationary orbit, meaning it stays above the same spot on the surface of the Earth at all times. The problem with this strategy is that to maintain geostationary orbit, satellites have to be very high—approximately 22,000 miles above sea level. That is far enough away that, even travelling at the speed of light, there is a noticeable lag for data to reach the satellite and then get back down to its destination on the ground. The other way to make satellites work is to have a lot of them in lower orbits. In lower orbits the satellites move much faster, so no single satellite will be over the same spot for very long. But if you have enough of them, you could get a signal to the closest one and have it relay the signal between a few more before beaming it down to its destination. Up until now, the main challenge for creating such a system has been the expense. Launching satellites, until recently, has been very expensive. You have to build a whole rocket for each launch. But with the Tesla/SpaceX creation of reusable rockets, building a network of a few thousand low-orbit satellites has become more feasible. Elon Musk may have cheered the successful rocket reentry for its potential in supplying the space station, but he was probably thinking instead about owning the entire internet.

Assuming that Elon Musk or Jeff Bezos will not encase the world in more than 10,000 satellites to own the internet, and, as nearly all the experts seem to think, that ground fiber is still the way to go, what can we do to coordinate democratic municipal ISPs? How can we free them from the grip of giant national fiber owners? Could we accomplish what Google is attempting, but public? Could we lay a democratically controlled and accountable public physical internet alongside the private physical internet currently connecting most homes on the continent?

It would be tough, but there are some regional examples worth examining. In Utah, the Utah Telecommunication Open Infrastructure Agency, or UTOPIA, is a consortium of sixteen cities working to build a fiber optic network that would reach every home and business across a significant swathe of the state. UTOPIA is a public enterprise, funded through government bonds, that is directly competing with the private owners of the physical internet. It makes use of peering agreements with other public and academic networks, thereby expanding its fee-free reach to internet locations around the country. But that reach is still very limited. UTOPIA owns the fiber in its area, but leases fiber to get to the major data centers outside its immediate area, data centers owned by yet more parties. And though UTOPIA peers with other publicly-minded networks, those networks themselves must pay to reach the privately-controlled bits of the internet. Maybe most interestingly, UTOPIA itself is not an ISP. That is, it provides people with the glass strands, but doesn’t code and carry their internet traffic along those strands. It doesn’t run the laser flashes that constitute the signals. Rather, approved private ISPs operate on UTOPIA’s physical network. UTOPIA calls itself an “open source” model, meaning that its physical network is available for use by any qualifying ISP, public or private.

The idea of UTOPIA is encouraging. That a public agency can viably fundraise and execute construction of a door-to-door fiber optic network, even if only regionally, is a small proof of concept. It could conceivably serve as a model for a national UTOPIA, a publicly-funded national network of fiber connecting local, regional, and state systems across the country. But UTOPIA’s history calls for some caution. The reason UTOPIA is not an ISP is that public entities in Utah are not allowed to be ISPs. And the reason they’re not allowed to be ISPs is that, after public internet efforts including UTOPIA initially got off the ground in Utah, the internet giants descended with near unlimited resources to fund lawsuits and lobbying efforts. They succeeded in effectively outlawing public internet service providers in Utah with legislation thanks to the American Legislative Exchange Council (yes, ALEC, the conservative legislative advocacy group responsible for a large percentage of the most private-profit-driven, scourge-of-the-left state laws across the country). 

The mere possibility of a public option for the internet in Utah led to a massive legal and lobbying effort that ended with a law banning public ISPs. The public broadband advocacy organization Community Networks offers an interactive map showing the country’s public internet networks, and notes that at least eighteen states have put up barriers to public internet. The millions of dollars spent by the likes of Comcast and AT&T to block net neutrality will be a drop in the bucket compared to the effort they will launch at every level of government to prevent any kind of nationwide network to shift power and leverage to local ISPs.

If we can’t build our own public, national, physical internet, is there a way we can force the current owners of the internet to act in the public’s interest? Can they be prevented from hamstringing local internet networks? This sort of move would have historical precedent. The railroads were notoriously built with public money for private profit, often involving sweetheart deals with associated or co-owned industries at the expense of everyone else. Much like Comcast carrying its own content for free but charging others, railroads would provide cheap or free transport to their subsidiaries but would charge others exorbitantly. Once the scandals reached a fever pitch in the 1880s, Congress designated the railroads as “common carriers,” meaning that, given their status as essential services, they were legally required to offer their services on a fair and equitable basis, and were to be closely scrutinized by regulators to ensure that was the case.

Congress did the same to the telecom industry in 1934. Rather than breaking up AT&T’s monopoly, or nationalizing the phone lines, Congress enacted a law to force AT&T to carry telephone transmissions from all customers on a fair and equal basis. Congress enacted a smaller but similar rule in the 1970s at the behest of the cable industry to restrict AT&T’s control over telephone poles. AT&T saw cable as a potential competitor both in terms of telecommunications and for its own plans to offer video delivery in the future. It also saw that the cable companies would have to spend a ton of money to run networks of wires if they couldn’t use the existing poles. So the cable companies found themselves being charged exorbitantly to hang cables on poles. The Pole Attachment Act directed the FCC to determine reasonable rates for pole owners to charge and thus set prices for pole leases.

The internet seems a natural choice for “common carrier” status. But, through decades of regulatory capture, lawsuits, and savvy lobbying, internet providers—both ISPs and large networks—have evaded designation as common carriers and maintained an almost wholly deregulated existence. Is now a moment for revisiting this situation? Nearly the entire American economy is wholly dependent on the internet right now. Those of us who are lucky enough to be working are doing so online. Public schools are holding Zoom classes and public officials are suggesting that children locked out of the internet at home (whether due to geography or poverty) go attend class in McDonald’s parking lots. The canard that broadband internet access is a luxury rather than a necessity, although long false, is growing more obviously ludicrous every day. These wires are our highways now. They are the paths we travel to get to work and school all day every day. Surely they are common carriers. 

Of course, the viability of this argument looks no better today than it did during the net neutrality fight in 2017. The White House, Congress, and the FCC are aligned in the strongest possible opposition to any internet regulation. And at the point we have the power to impose common carrier status, we should use that power to do better.


It would seem ridiculous to countenance a completely privatized system of roads and highways. Even with common carrier rules in place, imagine having to choose and pay for the use of various paths depending on where you’re going, what you’re carrying, what your means of transit are, etc. And imagine having to erect an entire bureaucracy to monitor the maintenance and conduct of the private companies in charge of the roads and highways. Maybe there are some libertarian readers salivating at this possibility, but I doubt it will appeal to most.

Is there a good reason to treat the internet differently? Sure, we built it in a different way, but that was due more to historical accident than any logical reason. Roads have always been recognized as essential for commerce, and their rights of way set aside for public use. The advent and prominence of cars came while the depredations of the privatized railroad system were still fresh, and federal and state governments were all too willing during the Great Depression and in the glut after World War Two to make massive public investments in direct road building (even while it would only subsidize private ownership of telephone lines). Politicians looking for rural support pushed through the USPS Rural Free Delivery program starting in the 1890s, which spurred the building of public postal roads reaching nearly every home in the country, at great public expense but to immeasurable long term benefit, so that people could get their mail delivered to their doors.

One potential rationale for public roads but private telecom is that the telecom system is more susceptible to innovation, and can benefit more from private competition. This hasn’t borne out for a whole host of reasons. AT&T operated as a government-sanctioned “natural monopoly” for decades, and after being forcefully broken up in the 1980s has consolidated nearly all of its former holdings back into itself. And, in fact, there hasn’t been that much innovation in telecom or internet hardware in the recent past. Fiber has been in wide use for decades and we haven’t come close to reaching its potential for bandwidth. There might be innovation to be had in the sending and transmitting equipment, which might be a colorable argument for private ISPs on a public fiber backbone, similar to the UTOPIA model. But even this might concede too much. As Vanessa A. Bee has written in Current Affairs and In These Times, there is little evidence that market competition is or has ever been necessary for technological innovation. The internet itself, after all, is a creation of public institutions.

What am I suggesting? Nationalize the wires. Rather than attempting to erect a whole new set of fibers to run alongside the existing fibers as a “public option” of sorts, we need to seize our soonest opportunity to take public ownership over the nationwide network of cables that is destined to continue serving as essential public infrastructure for the foreseeable future. And just as the federal government provided grants to improve and expand public infrastructure during the New Deal and post-war period, we need to provide funds for local and regional democratically accountable ISPs. This need not involve dismantling the existing ISPs. But a handful of private companies should not own the physical wires, and the means of access to those wires, and the content that flows along them all at once. 

(A note on privacy and security: A sane person may balk at the idea of government owning and controlling the internet. “The internet should be a radically democratic and subversive technology. They can already see too much. Do we really want to hand Stephen Miller and Palantir the skeleton key to every bit that travels the wires?” First, don’t kid yourself, the NSA is already in the private switching centers. Second, there is nothing more inherently trustworthy about Bank of America, Facebook, Google, AT&T, and Comcast than the government. But this is a legitimate concern. It is, however, a concern that can be dealt with using the right encryption protocols. Just in the past five years huge swaths of the internet have moved from HTTP to HTTPS, and end-to-end encryption has gone from the stuff of paranoid conspiracy theorists to an expected feature in communication apps. No security is perfect, and government-owned wires could still carry a risk of increased government surveillance, but Peter Thiel and Palantir will end up with your data as things stand now, and they will happily sell it to any government that will pay.) 

The history of nationalizing resources and industries in the United States is complicated. The official story is often that “we don’t do that.” But, of course, we do do that. We nationalized the paper currency industry with the creation of the Federal Reserve. We attempted to nationalize steel to break a strike in 1952. We nationalized passenger railroads in the 1970s. We nationalized most of the mortgage market in 2008 when the Department of Treasury took conservatorship over Fannie Mae and Freddie Mac. We also nationalized General Motors Acceptance Corporation in 2008-09, rebranded it as Ally Financial, and sold it back to private owners in pieces up until 2014. The common thread among these precedents is that we stepped in with public money when things were going very badly in order to subsidize private owners. Then, when things improved, we handed the assets back over so they could continue generating private profits.

These past examples are helpful precedents but they have things backwards. We should not have waited for a complete collapse of the banking system to nationalize paper currency. We should not have waited for a collapse in the housing market to take democratic control over Fannie Mae and Freddie Mac. So long as we use public money only to bail out private industry, we will forever be responding to suffering rather than preventing it. How many public school children are we going to send to the McDonald’s parking lot before we understand that internet access is a public good? In a better world, universal public internet access would be a political tool even more effective than the Rural Free Delivery Program. Politicians of all stripes and in all levels of government would see the golden opportunity presented by record dissatisfaction with private ISPs like Comcast, and would jump at the chance to institute a popular, useful program that would expand the economy and provide public works jobs. A champion of community internet would emerge, driving federal grants for municipal ISPs alongside federal and state programs to buy, build, or co-opt long distance transmission lines and switching centers.

If we have a strong understanding of what it means for something to be a public good, we will see the need to reach beyond defending those we already have—public education, the mail, the roads—and toward the public goods currently in private hands. The wires are a public good, necessary for education and governance and commerce. Whoever controls the physical internet controls a great deal of our lives, and our lives should be up to us. Freedom, as the saying goes, is participation in power. As the power of the internet grows, so must our participation grow. Isn’t that exactly what democracy is for? 


For further reading on the physical internet and its ownership and control, check out the following books:

  1. Infrastructure by Brian Hayes (2014 edition)
  2. Captive Audience by Susan Crawford
  3. Bit Tyrants by Rob Larson
  4. Mother Earth Mother Board by Neal Stephenson
  5. The Master Switch by Tim Wu
  6. Tubes by Andrew Blum

More In: Tech

Cover of latest issue of print magazine

Announcing Our Newest Issue

Featuring

A more punctual issue than usual! But no less full of surprises and delights.

The Latest From Current Affairs