Back when Google announced project Loon, their plan to provide Internet via floating balloons, I can remember thinking the idea was sort of crazy. It struck me as the kind of thing that might be useful in, say, parts of the developing world with little infrastructure, but here in the US, couldn’t they just somehow give us more Internet through the existing wires? I mean, the rate at which we can squeeze bits through fiber optic cables and wireless spectrum is undergoing the same type of regularly scheduled increase as the number of transistors that we can cram onto a sliver of silicon, so what’s the problem with getting everyone here in the first world connected?

Then I moved from San Francisco to Texas, and the longer I’m here the better I understand just how broken our current Internet access model is and the more I yearn for the day when the Internet balloons take up permanent residence above the hill country.

What follows is a collection of things I’ve learned first-hand, through the vehicle of my still-developing hobby WISP, about why American Internet access is bad and poised to get even worse.

Location, Location, Location

When my family and I first moved from San Francisco to Texas, we landed in the heart of Austin, and specifically in one of the first Google Fiberhoods in the city. We were there for about two years, but then this past summer we moved out to a multi-acre property about 30 minutes northeast of the city. Our land is surrounded by nearby cell towers, but there is no fixed-line ISP that reaches us or anyone within a few square miles of us in certain directions. As you drive by the homes and ranches, you see attached to the roofs of houses the universal, unmistakeable sign of bandwidth poverty – the HugesNet satellite dish.

My house has a dish, too, because, like the rest of the area, the previous owners relied on a mix of cellular and satellite service to get online. But cellular and satellite both suffer from meager bandwidth caps and rapacious overage fees, and in today’s world of 9MB web pages my neighbors and I routinely pay hundreds of dollars a month just to surf the web and do email.

Shortly after moving here I began looking into fixed wireless, but there are only three WISPs in our area and all them have one star on all of the relevant review sites; tales of two-year contracts and multi-week outages with no refunds are common. These companies’ plans also stink – you’re promised some 3 to 5Mbps a month for about $80, which is highway robbery, but you’re lucky if you get even that. At these rates and for such slow Internet, the though of just ordering a few Karma boxes for the house began to appeal to me. (One of the many nearby towers is Sprint, and Karma runs on the Sprint network.)

Given the unpalatability of our broadband options, I began researching wireless hardware immediately after the move. My property is up near the top of a hill, with very good views of much of the surrounding countryside. I figured that if I could find someone nearby who could get some sort of fixed-line broadband, then I could make an arrangment with them to beam bandwidth over to my place. I also considered renting space on a nearby cell tower and recouping the cost by selling the bandwidth to the neighbors.

It’s when I began seriously working on the cell tower option that the beauty and inevitability of the balloon idea truly clicked for me. Don’t get me wrong – on its face the Internet balloon idea is still actually nuts, but it’s vastly better than the terrestrial alternatives, which are all terrible because of one fatal problem: any moderately long-range wireless bandwidth solution needs line-of-sight to work, which means you need elevation, and (absent a viable balloon technology) elevation is a function of geography, which makes it eminently exploitable by rent-seekers.

This geography-based rent-seeking, by both wireless carriers (via national tower networks) and fixed-line ISPs (via cable, fiber, phone networks, and trunk locations) is the only reason that bandwidth quotas exist in 2016, and it’s what brings the Internet-of-balloons idea down from cloud cuckooland and into the realm of inevitability.

Yes, it’s incredibly depressing that physical location matters just as much for the economics of bits and bytes as it does for the economics of brick-and-mortar retail. And this goes from depressing to infuriating when you realize just how cheap wireless bandwidth now is, if only you had a good place to put a radio.

Wireless Hardware is Cheap

I eventually established a bandwidth-sharing arrangement with a local RV park that has a 50Mbps business-class connection, and I’m in the process of point-to-pointing that out to the other properties along our road. More on that some other time, though. What I want to talk about now is how insanely inexpensive quality wireless hardware is right now.

I went with Ubiquiti hardware for my hobby WISP, partly because of the great reviews and wealth of online resources, but also because the pricing is nuts.

I can buy a 1Gbps wireless link for $1,000 an endpoint. I recently bought two 150Mbps Nanobeam M5 wireless links for $72 per endpoint from Amazon, and if I had used a little more Google I’d have spent $20 more to get the Nanobeam AC endpoints that are rated up to 450Mbps. (Oh well, I’ll use those for the next expansion.)

I’m not a network engineer, but I’ve managed to set up two separate point-to-point links, and they function flawlessly. The hardware and software are very good, and it’s amazing that with a bit of relevant know-how, a few hundred dollars, and an Amazon account I’ve been able spin up a fully functional hobby WISP that now serves two customers and counting.

The “and counting” part is not wishful thinking, either. The word is now out in the ‘hood about what I’ve been up to, and my phone is blowing up. When people hear that they can get 10 Mbps/month from me with no quotas and no contract for a mere $100/month, they’re like “SHUT UP AND TAKE MY MONEY” – that’s how bad things are.

But as many people as there are here that are desparate to give me $100/month for Internet that doesn’t even come within a factor of two of the FCC definition of “broadband”, it’s still hard for a hobby ISP like mine to make money. The high-performance outdoor radio links are so cheap that they’re essentially “free”, but everything else costs an arm and a leg.

Elevation is Expensive

The real costs for a WISP are in the infrastructure and personell. Specifically, if I’m going to expand much further then, as I said above, I need elevation, and elevation is expensive for a few reasons.

First, even if you’re putting up a small, unregistered 50-foot antenna tower on someone’s property, you’ll need to pour a foundation, construct a small building to secure your equipment and keep it out of the elements, and provide both power and backup power to the building.

Once you’ve spent a few thousand on the tower, then you need to rent a boom lift to get a person up there to do the installs and any maintenance.

And then there’s the matter of any rent you may have to pay the property owner as part of the arrangement.

Full-blown cell towers are an order of magnitude worse, because the up-front costs for one of those can run up to half a million by the time you’ve got site security, backup power, and city permits in place. Such towers also require an OSHA-certified tower climber to do any installs or maintenance on equipment.

Again, all of this expense and hassle is necessary just to get some cheap radios above the tree line.

Towers and Incumbents

At this point, you may be thinking: if building your own towers is so costly, then why not just rent space on an existing tower?

Because incumbents, that’s why.

Most cell towers are owned by one of two nationwide cell tower networks, and even if you can afford their rates these guys don’t really want to talk to you unless you’re buying space on twenty or so towers.

The local water towers are just as bad, as all the space on them is typically bought up by incumbent WISPs who may or may not have any actual hardware on them. (If you’re a WISP, it’s often worth buying the space just to keep upstarts out of your service area, even if you don’t throw a dish up there and offer actual service.)

Why are these incumbent WISPs so awful to deal with? Why do all WISPs, at least in Texas, have crappy websites, exploitative plans, stiff contracts, poor coverage, no support, and one-star ratings? My guess is that these negatives are symptoms of a deeper problem, which is that there just isn’t enough money in the “local WISP” well to support a better set of customer experiences, thanks to the aforementioned issues of geography, incumbency, and rent-seeking.

Help from Above

All of this brings us back to balloons, or drones as Facebook would have it. Whether it’s balloons, drones, or satellites, it seems like there must ultimately be an aerial solution to the bandwidth dilemma, because an airborne device is the only way to get around all of the rent-seeking and incumbency problems that are associated with tying bandwidth to geography.

Until the balloons arrive and save us, I’ll continue building out my own little hobby WISP with a point-to-point and point-to-multipoint model by just linking houses to each other, instead of trying to link everything to a central point. The downside of this approach, of course, is that if one link goes down, then everything downstream from it goes dark. It’s a support nightmare, and it won’t scale. Gahh… if only I could figure out a way to keep a balloon aloft…