Project Kuiper Finally Here

After several delays due to bad weather, Amazon’s Project Kuiper finally launched its first batch of low-orbit broadband satellites on April 28. The company says it is in communication with all 27 new satellites. Amazon used a United Launch Alliance (ULA) Atlas V rocket launched from Cape Canaveral.

Amazon is under pressure to get the launches going and has a commitment to the FCC to launch 1,618 satellites by July 2026. That’s half of Project Kuiper’s first satellite fleet of 3,232 satellites. The company has seven more scheduled launches using the Atlas V rocket, 38 launches scheduled for ULA’s larger Vulcan Centaur rocket, and 30 launches scheduled on Arianespace, Blue Origin, and SpaceX.

The company’s name refers to the Kuiper Belt, a region of the solar system that exists beyond the eight major planets. The Kuiper Belt is named after the late Dutch astronomer Gerard Kuiper, one of the fathers of modern planetary science.

Project Kuiper’s stated mission is to provide fast, affordable broadband to communities around the world that are unserved or underserved with broadband. The company says it will offer three sizes of customer terminals that will support three tiers of service at 100 Mbps, 400 Mbps, and 1 gigabit. No prices have been announced yet for the receivers. The industry is waiting to find out what Amazon means by affordable broadband. It will be interesting if Project Kuiper significantly undercuts the prices charged by Starlink.

Project Kuiper is currently part of the Devices and Services division of Amazon that is also the home for devices like the Kindle, Echo, Fire TV, eero, Ring, and others. The company is headquartered in Redwood, Washington, and the satellites are manufactured at nearby Kirkland, Washington. Amazon claims the factory will be able to crank out five new satellites per day.

Launching a satellite fleet isn’t just about satellites, and the company is building earth stations around the world to communicate with the satellites as well as sites that can track the telemetry and keep satellites operating in the right orbits.

The satellites will be parked in a narrow band between 590 and 630 kilometers (367 and 392 miles) above Earth.

It’s going to be interesting to see what a second major LEO satellite provider means for the U.S. broadband market. There is a movement currently underway to divert more of the $42.5 billion in BEAD grants to LEO satellites. It will be interesting to see how much of that might go to untested Project Kuiper.

There have been no announcements when Project Kuiper would begin offering service. Starlink began a beta test service in the U.S. in July 2020 when the company had 595 satellites. Starlink has 1,260 satellites when it began retail broadband service in early 2021.

Increased Attacks Against ISP Networks

Netscout recently released its latest Threat Intelligence Report that documents DDoS attacks in the second half of 2024. As has been the trend for many years, the largest target of DDoS attacks has been ISP networks. There were over 8.9 million DDoS attacks documented in the second half of last year, up 12.75% over the first half of the year.

The report shows that the largest number of DDoS attacks came against wired ISPS and carriers. In the second half of 2024, there was 524,445 attacks against wired network, with the second most attaches against computer infrastructure providers at 197,221. This is not hard to understand since attacks against an ISP is the easiest way to make an attack against a specific target.

The report shows that the average DDoS attack against wired networks lasted for 33 minutes and that the maximum attack for the period used 937 Gbps during the attack.

Wireless carriers saw 24,433 attacks in the second half of 2024, with an average duration of 86 minutes and the maximum size of the attack measured at 119 Gbps.

Netscout noted that DDoS attacks are now outpacing other kinds of cyberthreats. DDoS attacks are launched by sending huge amounts of requests for connection to a specific network, often at critical times. The purpose is to effectively knock the network out of service for a while.

The report notes that a lot of the attacks are politically motivated. Attacks against Israel were up 2,844 percent last year, and attacks against Georgia were up 1,478 percent. The report notes that DDoS attacks are not just a cybercriminal tool but have become a dominant tool in geopolitical intimidation.

The report notes some of the regional causes for attacks. For example, in Kenya, a lot of attacks were launched to protest finance bills. In Mexico, the attacks spiked to coincide with elections.

The attacks are often launched by linking low-power Internet of Things (IoT) botnets with high-performance enterprise servers and routers, which amplifies the efficiency of the attack. Netscout notes that artificial intelligence automation is helping to prolong the attacks. Attackers are using a variety of methods including carpet-bombing, IPv6 abuse, ISP masking, and geo-spoofing. Unfortunately, many of the tools used to fight back and end DDoS attacks are losing effectiveness.

FCC Loses the Ability to Levy Fines

On April 17, the U.S. Court of Appeals for the Fifth Circuit ruled that the FCC doesn’t have the authority to levy fines. The case under appeal came from an AT&T fine issued by the FCC in 2024 against the company for violating customer privacy by selling customer location data. AT&T appealed the FCC ruling and said that the FCC had violated the Seventh Amendment of the Constitution by not allowing AT&T to have a jury trial.

The Fifth Circuit sided with the AT&T appeal and said that the FCC had acted as prosecutor, judge, and jury in deciding that AT&T had broken the rules and in setting the fine.

The Court ruling was not particularly kind to AT&T and is worth a read. The Court recounted the facts of the case and said that AT&T was, at a minimum, guilty of negligence since it assumed that those who purchased its location data would protect consumer’s privacy rights.

The Court’s opinion relies in the 2023 Supreme Court Decision in SEC v Jarkesy that ruled that the SEC had no authority to levy civil fines. Since then, many lawyers have assumed that the ruling would apply broadly to all federal agencies, and this ruling affirms that opinion. The two cases together seem to affirm the inability of federal agencies to fine the companies they regulate.

This ruling will stand unless the FCC appeals it. FCC Commissioner Brendan Carr and Commissioner Nathan Simington both objected to the original fines against AT&T. However, the FCC has issued fines since 2023 against others, including a $4.5 million fine levied this year against Telnyx LLC related to making imposter robocalls leading up to last year’s elections. The FCC already wrote a letter to other courts that are reviewing cases involving FCC fines and asked those courts not automatically follow this decision.

The case also brought into question the process by which the FCC decides that its rules have been broken. The Court found that the FCC had essentially judged AT&T as guilty without giving the company the chance to defend itself. This might mean that the FCC and other regulatory agencies will be limited to gathering facts and forwarding cases to the DOJ without making any findings of wrongdoing. I don’t want this to sound like an overexaggeration but that philosophy seems like it would largely take the FCC out of the regulatory business. It would mean that the agency can set rules but can’t decide that a regulated company has broken its rules, since doing so would be a finding of wrongdoing without a trial.

I’m not a lawyer, and there are nuances to this ruling that I am probably missing. But this doesn’t seem to bode well for the FCC and many other regulatory agencies. The ruling would seem to curtail the ability of Chairman Carr to pursue a lot of the topics on his 2025 wish list.

Technology Equality

There was an article published last week by Dr. Christopher Ali in Tech Policy.Press that asks if we should be making widespread broadband grants to Starlink and other low-orbit satellite technologies. Dr. Ali is Professor of Telecommunications in the Bellisario College of Communications at Penn State.

I highly recommend reading his paper. I was particularly taken by his conclusion. He says, “There is an important difference between technological neutrality and technological equality. LEO and fiber are not equal, and any policy that treats them as such will widen the very divide we have spent decades trying to bridge.”

I have been making this same observation about almost every grant program in the last decade, but just not as eloquently as Ali. As Ali points out, his comments are not a criticism of Starlink. Like him, I’ve talked to dozens of rural folks who absolutely rave about Starlink. For rural households, finally getting access to working broadband twenty years after the rest of the country has been a revelation. They can finally work at home, join Teams calls, and take online classes – things that the vast majority of Americans take for granted. Starlink should absolutely be a part of BEAD to reach remote locations, but should it be deployed to other locations?

Ali’s real issue is with the folks who set grant rules. Consider BEAD. The rules were established in 2020, and grants were not expected then to be finally completed until the end of 2028. Consider how much broadband has changed in the country just between 2020 and 2025.

  • In 2020, Ookla said the median broadband speed in the country was 86 Mbps download and 12 Mbps upload. In March of this year, Ookla says that median broadband speeds in the country has increased to 287 Mbps download and 53 Mbps upload.
  • In the second quarter of 2020, OpenVault said the average U.S. household used 359 gigabytes per month of download data and 25 gigabytes of upload data. OpenVault says at the end of 2024 that consumption had grown to 652 gigabytes of download and 46 gigabytes of upload.

The policy folks who set the BEAD rules set the broadband target performance for BEAD just a hair above the national average broadband performance in 2020. We’re only half way to the completion of BEAD grant construction and the country has already more than doubled the 2020 national broadband speeds and consumption. It’s not a stretch to predict that by 2028 the average U.S. home will be consuming more than a terabyte of data each month.

If the authors of the BEAD grant rules had looked just a decade forward, they would have set the BEAD performance standard to something like 400/100 Mbps. It doesn’t seem like a big policy stretch to think that valuable grant money ought to build networks that match the average market performance when they are completed. As Ali mentions, the biggest issue with LEO satellites isn’t even speeds, but capacity. Will the LEO companies be able to provide broadband to the many millions of households who will have no other broadband options?

It’s obvious why the folks in Congress picked wimpy BEAD standards. They are politicians and were under tremendous pressure from ISPs to not be excluded from BEAD dollars – and even under more pressure to not declare cable company networks as underserved. I remember the furor from cable companies in 2020 that lobbied hard against the BEAD upload speed requirement of 20 Mbps. That was because, at that time, most of them had upload speeds closer to 10 Mbps. It’s amazing what only five years of market pressure has done, and cable companies are upgrading urban upload speeds to 100+ Mbps with quick mid-split upgrades and have plans to get to gigabit upload speeds with DOCSIS 4.0.

WISPs didn’t have the same market power as cable companies in 2020, but they fought hard to make sure that the requirement for BEAD didn’t climb above 100 Mbps. But after only five years, they got access to a lot of new spectrum and can buy gear that will deliver 500 Mbps or faster broadband.

As ALI points out, LEO technology barely meets the 2020 definition of broadband that was codified in BEAD, and it is not a forward-looking technology – it is not equal to fiber or even to fixed wireless. And yet, the NTIA is doing mental gymnastics using an argument about technology neutrality to give more money to satellite technology. Perhaps the critics of satellite technology will be proven wrong, and satellite providers will improve technology so that by 2028, they will be delivering forward-looking speeds and coverage. But if not, we’ll be making grant awards in 2025 to implement 2020 broadband.

Four Trends in Infrastructure Spending

There is an interesting article from the Brookings Institute that documents four trends in infrastructure funding. The conclusions of the report surprised me and I suspect they will surprise others.

Infrastructure Spending is Up, But Spending as a Share of GDP is Down  

Total infrastructure spending has been climbing steadily over the last fifty years. Infrastructure in the early 1970 was a little over $300 billion per year, and since 2015 has been between $600 and $640 billion per year. However, the comparison of infrastructure spending to gross domestic product shows a slight decline over these same fifty years. In the early 1970s, infrastructure was 2.5% to 2.7% of GDP and in the 2020’s has dropped to 2.2% to 2.3% of GDP.

Gross domestic product is the measure of a country’s total economic output, expressed as the monetary value of all goods and services produced. The ratio of infrastructure spending to GDP shows the amount of national wealth that is reinvested in infrastructure. Seeing this ratio drop in recent years is an indication that, over time, the country is not reinvesting at the same level as in the past. Brookings says that most of the spending drop is related to highways, with the other categories of infrastructure spending the same or higher over time.

State and Local Governments Fund Most Infrastructure Spending

This is the finding that surprised me the most. The percentage of infrastructure spending that is funded by State and local governments has climbed steadily since the 1970s. In the 1970s, State and local government spending paid for 67.9% of infrastructure, and that has climbed to 77.5% since 2020. The percentage of federal spending has dropped even with what feels like huge infrastructure spending from the Infrastructure Investments Job Act and other federal programs.

To show the magnitude of the shift, in the 1970s, the federal government spent $1 trillion on infrastructure while State and local governments spent $2.2 trillion. In the 2010s, federal spending had increased to $1.3 trillion while state and local spending more than doubled to $4.6 trillion.

The Share of Infrastructure Spending on Maintenance is Climbing

The percentage of infrastructure spending to maintain existing infrastructure is climbing, meaning the percentage of spending for new infrastructure is dropping. In the 1970s, 44% of infrastructure spending was to maintain assets like roads and buildings, and that has grown to 57% since 2020.

State And Local Revenues More Important than Federal Grants

State and local revenues are paying for an ever-increasing percentage of infrastructure spending. This means state and local taxes and user fees. In the 1970s, 59.5% of infrastructure was funded by local taxes and fees and that’s up to 71.6% in 2023. This is also a surprise because of IIJA funding, and Brookings speculates that increased federal grants drive State and local government to fund more money than normal to use as grant matching funds.

John Deere’s Private 5G Network

FierceNetwork recently published an article talking about the private 5G network implemented by John Deere for it’s campuses and factories. Having a secure private wireless network is allowing the company to implement 5G technology onto the manufacturing floor.

The company got started on this solution when it spent $546,000 to buy CBRS PALs licenses in 2021 in five counties in Illinois and Iowa, that included its headquarters in Rock Island County, Illinois. John Deere has more than twenty facilities in the five counties.

The key to making this work was the exclusive use of spectrum, meaning that Deere can always rely on the wireless bandwidth free of interference or use by others. That assurance allows the company to establish a wireless network to communicate directly with manufacturing equipment in its factories.

It’s taken four years for the company to make the transition to wireless broadband, with the company having to find a 5G interface for each piece of equipment. Jason Wallin of John Deere is quoted that the company adopted an 80-10-10 model where 80% of broadband needs are covered by the private 5G networks, 10% by WiFi, and 10% with hard-wired Ethernet. The company says it is spending $1.7 million this year on updated smart tools to take advantage of the wireless network.

The company was quoted in 2021 about seeking the same wireless solution in its other factories around the world. But adopting this in other locations would require working with a cellular license holder to somehow carve out spectrum for each factory.

The creation of 5G private networks was one of the original big promises for 5G. There have been some notable examples of private 5G networks created around the world for Airbus, Shanxi Coking Coal, Group, Tesla, Lufthansa Technik, and NEC. However, this has not been the booming business for cellular carriers they hoped for. If you look back at industry press five years ago, you’d see the carriers touting this as a major opportunity. However, like other business lines based on 5G, the performance badly underperformed the expectations. The John Deere example showed a corporation that found it better to buy the spectrum and tackle this on its own rather than partnering with a carrier.

The CTO of John Deere says the company expects to use the newly created wireless network far into the future. This sounds like a relatively small investment to gain the benefits of fully automating the factory floor by connecting all machinery to an integrated network.

Eliminating FCC Regulations

I’m sure nobody is surprised that businesses regulated by the FCC responded with long lists of regulations that should be eliminated after the FCC invited them to do so in the Delete, Delete, Delete proceeding. Seemingly every trade association has an existing list of hated regulations, and the FCC heard from telcos, cable companies, cellular carriers, programmers, satellite companies, and even more obscure parts of the industry like prison phone providers and robocallers.

Some of the requests were fully anticipated and had been mentioned in the Delete, Delete, Delete docket. For example, AT&T and Verizon asked the FCC to eliminate most of the rules related to maintaining or reporting on copper telephone networks. The companies are tearing down those networks and would like to finish the task without paperwork or having to notify the public.

The purpose of the Delete, Delete, Delete docket is to eliminate rules that are antiquated and no longer needed. Trade groups took this as an opportunity to ask the FCC to eliminate regulations across the board, including regulatory rules that were explicitly ordered by Congress or rules recently adopted by the FCC. For example, there were multiple requests from ISPs to eliminate broadband labels, which just went into effect last year. Prison phone companies asked to overturn a recent FCC order that reduced prison calling rates. Verizon asked the FCC to reverse a recent decision that requires it to unlock phones so that customers can more easily change carriers. AT&T asked the FCC to eliminate its authority to issue fines after the company was recently fined.

One of the more interesting requests came from debt collection agencies and those who represent robocallers. They want to get rid of the rules that let the public easily opt out of receiving robocalls or texts. This doesn’t seem likely since the FCC is still actively working to eliminate junk and spam calls.

It is going to be very interesting to see what the FCC does with the huge pile of requests they received. I don’t know if anybody has added up the industry requests, but it wouldn’t be surprising if there are 10,000 specific regulations somebody wants to kill. The normal process in the past to add or delete regulations has been to issue a list of the specific changes being considered to get industry feedback. I assume that FCC staff will have to somehow wade through the huge list of requests to find those that meet the spirit of Delete, Delete, Delete. A lot of what is being suggested by the various industries are changes to active regulations that are not obsolete.

This is going to be a test of how activist the new FCC is going to be. If the Commission sticks to the stated purpose of eliminating unneeded regulations, this docket will not likely cause a ripple in the industry. I think everybody agrees that federal regulators should periodically eliminate outdated regulations.

However, the FCC could use this docket as a short-cut way to eliminate piles of active regulations. If that happens, it seems likely we’ll see scads of lawsuits challenging the FCC for not following the prescribed process for making regulatory changes. Interestingly, some recent court orders question the authority of regulatory agencies like the FCC to add new regulations that were not directly by Congress, and it seems to me that this same concept would apply to eliminating regulations.

Cleaning Up After Construction

I recently read an article from the ABC TV station in Cleveland that covers local complaints about damage caused during fiber construction. It’s titled. “It’s Terrible!: Local Communities Angered about Damages Caused by Broadband Installations”. This is not a headline that any ISP building a network wants to see.

In this particular case, the construction was being done in Brunswick Hills Township by Frontier, who used MasTec as the construction vendor. The story includes a complaint from a resident who said MasTec had hit a buried electric line and that she was reduced to using candles for months. The construction also hit a gas line and recently hit another buried electric line that knocked out power for 200 for ten hours.

Anybody who has ever buried fiber is familiar with these kinds of incidents. Sometimes the problems are caused by operator error where the company doing the construction makes a mistake. However, the problem often comes from the locating vendors that misidentify existing buried utilities. Folks who bury fiber would love to never hit any utility, but it unfortunately happens with fair regularity.

More troublesome in this story is a claim that the construction contractor left a large open hole behind them. It’s worth watching the video, because there is a huge hole that looks like it was intended for a fiber vault. The hole looks to be well off what normally be considered as public right-of-way where construction is allowed to dig without an easement from a property owner.

The real issue of the story is that Frontier and its contractor left behind some big messes. This story must be concerning to the many construction contractors that get this right. Contractors generally have a person assigned to construction projects to deal with the problems that inevitably arise. This particular incident would never have been a story if the situation had been dealt with immediately to the satisfaction of homeowners.

Let’s face it. Fiber construction is messy. Boring means digging holes, which quickly become messy in wet weather. Construction in the right-of-way upsets homeowners who think of the public areas next to the street as part of their front yard.

A lot of my ISP clients like to get ahead of these issues. They put doorhangers on every home in a construction area a week or so before construction to tell them what to expect. Most ISPs insist that contractors return areas to as near to original condition as possible after the construction. Most contractors today take pictures of areas before and after they dig as proof that they did the restoration work after digging.

ISPs also often talk to town officials ahead of time to tell them what to expect, and to give them somebody to contact when they hear about problems. Nobody wants a town official on TV complaining about them like happened in this story.

An ISP that makes a mess also has to worry about resident who refuse to buy service from them. A lot of my clients look at the construction process as a sales opportunity, and they knock on doors during the process to apologize for the commotion and mess and to introduce their new fiber broadband.

Broadband is Critical Infrastructure

The Kansas legislature recently enacted HB 2061 that declared that broadband is critical infrastructure. Specifically, the new law says that it is a felony to trespass or damage aboveground and belowground lines, cable and wires used for telecommunications or video services. When Governor Laura Kelly signed the bill she said it critical that all Kansans have access to reliable high-speed internet.

Kansas joins Florida, North Carolina, South Carolina, and Tennessee in classifying broadband networks as critical infrastructure and making it a crime to interfere with or damage telecom networks.

At the federal level, the Cybersecurity Infrastructure and Security Agency (CISA) includes communications technology as one of its sectors of critical infrastructure. The CISA website says, “The Information Technology Sector is central to the nation’s security, economy, public health, and safety, as businesses, governments, academia, and private citizens are increasingly dependent on its functions”.

This might seem like feel-good legislation sponsored by the telecom industry. However, the real goal is to make it easier for law enforcement to go after those who vandalize networks. There is a notable uptick in reported theft and damage to our networks. A whitepaper by NCTA, CTA, USTelecom, and the Rural Broadband Association highlighted a nationwide survey of telecom providers for June through August 2024 that identified 3,929 incidents of theft and vandalism, which is 44 incidents per day. The attack on networks affected service to more than 325,000 customers. In addition to knocking neighborhoods out of service, the attacks have disrupted 911 centers, emergency responses, and hospitals.

The current primary driver of damage to networks is copper thefts. But there are also numerous reported attacks on utility poles, cellular towers, telecom construction projects, and fiber optic wiring. The attacks aren’t limited to telecom networks, and there is a growing number of attacks on electric grids and EV charging stations.

The Kansas law hopefully makes it easier to penalize those caught in vandalism to networks. This law is part of a national strategy to make it easier for law enforcement to work with carriers to reduce vandalism. Raising crimes against networks to felonies means a higher priority to pursue those doing the damage, and just maybe, a little less desire to do the damage.

What’s Your Broadband Journey?

Anybody who was using computers before 2000 can describe a broadband journey of how they communicated over the years. That journey mostly describes the broadband choices we each had in our neighborhood.

My broadband journey started sometime in the early 1980s. I had a TRS80 home computer from Radio Shack. I played around with programming and learned the basics of computing. But I  really started to enjoy the computer when I discovered bulletin boards. I lived in Oakland, California, at the time, and there were hundreds of different bulletin boards available with a local call. These were text-based forums where you could type messages (punctuation often not allowed). I bought what I recall as a 300-baud modem that let me connect to different boards. I spent a lot of time on music, science fiction, and sports boards. About the same time, I joined CompuServe. I didn’t like the chat rooms there as much as I did the bulletin boards, but it was my first online introduction to being able to follow the news.

Over time, I upgraded to better computers. I eventually joined AOL and used it mostly for email and to play online trivia games. Dial-up modems by this point were up to 56 Kbps, a blazing speed compared to my first modem.

Since I was in the telecom industry, and I closely followed the development of DSL. I had moved to Maryland, and I recall buying DSL service from Verizon in late 2000 at a blazing speed of 1 Mbps. My office at the time was served by a T1 at 1.6 Mbps, shared by a dozen people. I recall being amazed that I could get 1 Mbps at home just for me. DSL was so much faster than dial-up that it felt like it ought to last for many years.

But it didn’t last. AOL and other web companies piled the web full of pictures, and 1 Mbps soon started feeling slow. A few years after getting DSL, I upgraded to a cable modem from Comcast. My recollection is that it delivered 3 Mbps download.

Verizon announced that it was building fiber in the DC suburbs, and I got on the waiting list years before the product launched. My neighborhood was one of the first neighborhoods to get FiOS fiber in Prince Georges County. I was the first customer in the county that Verizon tried to connect with FiOS voice and the Verizon installation crew was at my house for almost three days in 2004 trying to get voice to work. My FiOS connection gave me 30/5 Mbps, which felt like lightning at the time. At the time, I couldn’t figure out the benefit of having 5 Mbps upload, when most people didn’t even have 5 Mbps download.

My next move was a step backward. I moved to St. Croix in the Virgin Islands. I scoped broadband before I went. The telephone copper network was so poor there that voice barely worked, and the cable company didn’t yet offer broadband. I subscribed to Broadband VI, a local WISP, and I got a solid connection because I bought a home in the shadow of one of their towers. I was on this service for almost a decade. It was reliable, and my average download speed at the end was around 5 Mbps – fast enough for working from home at the time.

When I moved back to the States, I bought a house that didn’t have broadband, but I wasn’t worried since Comcast had a pedestal at the end of my driveway. In a story that I am sure is familiar to many, Comcast customer service said my house couldn’t be served, and it took a few week of calls to convince them otherwise. Comcast wouldn’t sell me standalone broadband and forced me to buy a basic video package – the cable box went into the closet and was never used. My recollection is that my speed was 60 Mbps, and I got a notice after a while that speeds were unilaterally going to be upgraded to 100 Mbps.

I moved to Asheville, North Carolina and got Charter broadband – no other bundled products were required. My initial speed was 100 Mbps, but it eventually got upgraded to 200 Mbps, and we upgraded to 700 Mbps to try to get better WiFi distribution in our hundred-year old home.

Of all of the broadband products I’ve purchased, I think the 1 Mbps DSL modem from Verizon was my favorite. It was 30 time faster than dial-up and, for the first time, I felt freed from the limits of the connection technology. At the time, that was enough speed to do anything I wanted. I’m also still nostalgic about the beeps and boops of my first modem.

Feel free to share your broadband journey in the comments.