Starlink ocean coverage without inter-satellite links?

While it would be easy for Starlink to cover continental areas and essentially area where suitable land with fiber-optical cable near by, the question is whether the current Starlink technology can cover the whole world? Since 70% of the Earth is covered by the water, could Starlink provide universal coverage there too?

Even at 550km altitude, each satellite covers a circle with almost 2000km diameter. That is no small feat, since it takes only 20 satellites to cover the whole area on a single orbital plane. Real issue is how to place gateways around the world where Starlink satellites could route the traffic to the Internet. 

As I explained in previous post, placing the gateways on the continental edges and islands can cover most of the ocean areas, especially in the Atlantic. But there are some gaps that can be addressed by using gateway platforms stationed in the middle of the ocean. Since SpaceX already operates three autonomous spaceport drone ships, (ASDS), they could easily use the same platform as mobile autonomous gateways capable of operating in the middle of the ocean. Estimated cost of ASDS is 10 million USD, and annualized cost per ten years of operation is around 4 million per ASDS. It would probably be less for pure gateway platform, due to less frequent "travel". 

Any island with fiber optical link could be used as a gateway station rather cheaply. 

How many ASDS stations would be required to cover Indian ocean? Three? Additional ten for the Pacific? Some in South Atlantic? And a few spares? With a fleet of 20 drone-ships SpaceX could provide coverage of all major oceans. That is 200 million investment. Ok, the buildout might take some time, but it shows that Starlink can cover majority of oceans even in its present form, without additional inter-satellite optical links. 

This is dirt cheap. Viasat-2, higher performing GEO high throughput satellite, costs three times as much. Viasat-3 constellation will cost six times as much. Also SES mPOWER. 

Sure, such ship based gateways will have additional operational costs (station keeping will require frequent refueling), but for a fleet of 20 ASDS, that should lead to less than 2 million per year. But that would essentially kill most of revenue potential of traditional GEO communication providers, such as Inmarsat. 

Having a fleet of ASDS seems like a nice proposition. But then again, the best case would be to have no such fleet of ASDS. You need to tug them. Refuel them. Bring back to port for repairs and refitting. There are going to be outages. The best would be to not to take care of such large fleet.

Here the large number of satellites comes handy. Each Starlink terminal can reach between two and four satellites. But why would each ship based terminal have just one antenna? If it has two, each ship can act as intermediate gateway between two satellites. And there is a large selection of ships that can act as intermediate gateways.

Snapshot of AIS enabled vessels from marinetraffic.com

Maritime satellite communications market is almost 3 billion USD per year. For example, InmarSat BGAN (Broadband Global Area Network) designed for maritime applications offers whopping 492kbit/s. And terminal costs 3-10 times more that Starlink user terminal.  VSAT costs are higher, with average revenue per user (ARPU) of over 200USD per subscriber. Something that SpaceX can easily match and undercut with the superior service. 

By using dual terminals as intermediate gateways, Starlink could easily cover the gaps in the oceans, but at the expense of network delay (that could potentially double or even tripple). But that would still stay below 100ms round trip, far better than any GEO latency. Another drawback would be lower effective utilization of the available frequency bands, since each bit would be transmitted to space and back at least twice. But that is not a big drawback, since Starlink will not likely to be fully utilized over the ocean anyway. 

Even more lucrative would be aviation broadband market. Total In-flight connectivity is estimated at five billion USD (although due to COVID-19 in 2020 it will be just 1.5 billion). Starlink would be especially well suited to address long-haul transcontinental flights, which could easily "piggy-back" on intermediate maritime gateways. 

Snapshot of world air traffic during COVID-19 from flightradar24.com

Why not providing continental support for aviation broadband? From technical perspective, Starlink can provide this even today. But there are legal and business considerations why this will not be the priority. First is related to local licensing issues, since SpaceX would require getting approvals in each country addressed. Once operating licenses are obtained, it might be more valuable to offer fixed land subscriber connectivity first. Simply due to fact that they will be able to better utilize their frequency spectrum and subscriber density by using much larger oversubscription for individual subscribers than for aviation industry. But over the oceans and international waters, there would be less subscribers and less legal framework to address. 

Nothing stops OneWeb, Telesat LEO or mPOWER to provide competing service for maritime and aviation applications. But just like Tesla in automotive industry, real threat here is vertical integration. SpaceX controls terminal design, production, ASDS operations, satellite production, gateway deployment, launch and the network management inhouse. All other competitors are essentially specialists for a single segment - satellite operations and network management. All other parts are outsourced. Even without any technology breakthroughs, without Starships, without inter-satellite links, Starlink is already in position to wipe out possibilities of growth for other space communication operators. With truly global coverage available and low cost terminals, they can address "safe-havens" of other satellite operators - mobility, government, enterprise, maritime and aviation markets with much better value proposition. How can established satellite operators fight against Starlink? Their most effective weapon is established positions in various markets, frequency licenses and government relations worldwide. Thus we can expect even stronger campaign against Starlink through various governmental in intergovernmental bodies in order to slow down Starlink progress.