Starlink major victory - RDOF low latency eligibility

Now that was a surprise. FCC just published that they accepted SpaceX explanation and allowed LEO constellations to bid for low latency subsidy in the upcoming RDOF auction. 

This is huge. This is big deal. This makes Starlink solvent. This makes LEO broadband a reality. This makes space really a new economic frontier. Just consider the numbers. Add Viasat, SES, Hughes Network Systems, Inmarsat, Eutelsat, Intelsat, Telesat together. Their combined revenue is around than 12 billion USD. In US only, satellite based broadband has two million subscribers, with ARPU edging over 80 USD. 

But RDOF auction enables SpaceX to easily outbid any other satellite broadband provider, due to low latency tier advantage. Essentially, they can get almost all the money. 16 billion USD. Over ten year period. The only thing they need to do is to build a gateway with high speed fibre cable Internet connection up to 500 km away from the bidding area. No part of continental US is that far from decent fibre cable connectivity. Plus they can cover Alaska. Only remote islands might be out of reach (until inter-satellite links become reality). And all other LEO providers cannot be eligible. OneWeb is in the bankruptcy. Plus their cost structure is much worse than Starlink. 03b and mPOWER are in the same tier as geostationary satellite broadband providers. Net result? Starlink will probably be awarded majority of RDOF funds. Let me be conservative. 10 billion. Payments spread over 10 years. That is guarantied revenue of a billion USD PER YEAR. Without any paying subscribers. Without any military contract. Without any in-flight connectivity contract. Without shipping customers. Or mining customers. Without any subscriber from any other country in the world. 

There is a caveat: SpaceX needs to prove until the submitting date that it is capable of providing Internet connectivity with latency below 100ms. Impossible? It was reported that in recent Reddit AMA for SpaceX software developers that they transfer 5TB of (telemetry) data every day. That is 463Mbit/s on average. Ok, they probably have many separate terminals under the test. But their current gateway map has 28 stations in the continental US. It is not clear how many were operational. And their minimal coverage constellation is just under the 50% complete. We don't know how many user terminals are in use.

It is important to note - they do not have to prove seamless handover from satellite to satellite. They do not have to prove continuous aggregate bandwidth. They just have to prove that packet round trip time (RTT) is less than 100 ms. 20 ms is the additional delay introduced by physics. 80 ms is the delay that most network engineers can easily address. Unless they are really incompetent. And if SpaceX has shown anything, then it is that they are really competent engineers. 

In mobile networks, distance between user terminal and the cell will be less than 5km. That is a hundred time less than for Starlink. So any broadcast time plan sent from cell to connected mobile devices  can be updated 100 times faster. So a mobile cell will be able to allocate the required radio broadcast slots much faster than for Starlink. That represents a problem for Starlink. To achieve desired RTT, Starlink has to allocate some time slots to each connected terminal whether they are used or not. Lets say they need to allocate one Ethernet packet size (1500 bytes) in each timeslot (lets say 20ms) for each connected user terminal. Lets say 500 terminals per satellite. That leads to 300Mbits/s of upstream capacity per satellite, whether it is used or not. Or 600kbit/s per user terminal. Or 300MHz of uplink spectrum. In this example, I have avoided compression, error correction, and various modulations that could increase or decrease frequency utilization. 

Of course, Starlink could offer cheaper, higher latency plan and more expensive, low latency plan. But there is no fundamental issue why they could not demonstrate low latency Internet connection. It is enough that demo terminals receive reserved bandwidth/uplink time slots and they could easily achieve desired latency. Only thing that could affect them if they optimized their network for extremely high frequency utilization. But FCC/RDOF rules have been known for a long time. And SpaceX definitely expected to bid for low latency tier. The real news here is that they are allowed to bid IF they can demonstrate it before the submission deadline. The real issue is that it might force them to uncover user terminal operation (in public beta), before they feel ready. Real issue would be to demonstrate half baked system that might crash often, or have issues with satellite handover. But to demonstrate less than 100 ms round trip latency? I believe SpaceX could demonstrate that capability even for initial test. On day one. 

So why is FCC skeptical? One problem inherent to radio networks in many to one topology is avoiding interference caused by multiple terminals emitting on the same frequency at the same time. To efficiently use the available spectrum, central locations typically define a broadcast emitting plan, that is pushed from the central location (gateway) to the connected terminals. That plan precisely defines which terminal will broadcast on which frequency on which timeslot. This is not unique to satellite networks. The same applies to mobile devices connecting to mobile network cells. That is why mobile networks cannot function without accurate timing information provided by GPS.

Another reason for FCC skepticism might be face covering. They say: we doubt any LEO constellation could do this. It is not proven. Unless this is demonstrated. And SpaceX is the only one that can demonstrate it until 15th of July. OneWeb could too, but they are in the Chapter 11 process. So it is not likely they will have resources to demonstrate the same. And no-one else can show actual, real time demonstration. Which means...almost no-one can compete with low latency last mile costs that SpaceX has. Everybody else has to invest additional money for any county they wish to support. SpaceX can leverage any investment over all areas covered by Starlink. And provide low latency. So FCC will have no option but to accept Starlink demonstration. Imagine political consequences if they don't. They refuse to accept Starlink low-latency bid and six months later SpaceX rolls out Starlink nationwide. And they will sue FCC. And somebody in FCC will lose his job when the government changes. Because they might be procedurally right, but they will be easily proven technically wrong. Maybe even before they pay out the first subsidy. Government will pay 16 billion dollars to anybody but the company that provides the desired service nationwide? FCC leadership does not want to look stupid. 

I am sure they have considered what has happened with USAF block buy of ULA rockets. Or in Alameda county. Elon Musk has growing political power and influence. He has 20 million twitter followers. Close to a million US citizens are driving Tesla cars. And they love it. And the number grows every day. He is the beacon of entrepreneurial spirit, a true embodiment of American dream. Rags to riches story of self made billionaire. Someone who is really standing behind his beliefs. A true entrepreneur superstar. On par with Bill Gates. Thomas Alva Edison. Steve Jobs. Howard Hughes. John D. Rockefeller. George Westinghouse. Henry Ford. And probably even greater.

In essence, US government just funded Starlink build-out. And who can compete? Amazon? Really? They will invest 10 billion? Over 5 years. While their main competitor is earning billions? Not even Amazon is that strong, starting five years late, without ten billion subsidy and entering a saturated market? 

Who can compete? Not Amazon. SES/Viasat/Hughes? Yes, but they don't have subsidies. They don't have economy of scale. They don't have very low launch cost. Nor mass production of satellites. But they do have the know-how. They have existing subscribers. They have remote gateways. 

So FCC effectively selected Starlink as the future of worldwide satellite broadband. New monopoly, that will make AT&T seem small in comparison. Other world powers might not like it. They will have to subsidize their own. Which might, just might, resurrect OneWeb from ashes. 

I think this FCC decision will affect the future of space exploration more than Commercial Crew. More than Artemis. It makes Starlink viable. And viable Starlink gives launch demand for Falcon 9. That is one order of magnitude cost reduction to LEO. And Starship will be another. Because this FCC decision gives Starship a commercial purpose. An anchor customer. 

I wonder if FCC leadership was aware of the consequences of their decision. This could literary spread human civilization to the other planet. It is a pivotal moment, equal in consequences to the COTS contract award to SpaceX in 2008. Without that, SpaceX would probably be a minor player. Or it would cease to exist. There would be no Falcon 9. No Falcon Heavy. No Commercial Crew. No rocket landings. Geosat launch would not cost 50 million. More like 150 million. Commercial Crew mission would cost half a billion. There would be no Artemis. Because SLS would cost 2 billion. And it would not have any alternative. Without Starlink, there would be no Starship. And with low cost, low latency Internet connection, life will get much better in remote areas. With solar panels, batteries, and high speed Internet living in a remote tropical paradise might become a new normal. Off-grid will become equally "cheap" as being in urban areas. And quality of life will improve. Greatly. This single decision might become equally important as the mobile phone itself. Or smartphone. Connected to humanity. Anywhere on the planet. 

It is a brave new world indeed. And this decision will make it happen. It will be disputed. Considered as subsidy. Free money. Unfair. Uncompetitive. USG conspiracy to rule the world. But just like the geostationary satellites, Internet, GPS, smartphone...it will change the world. And make mass space travel a reality.