Autonomous Vehicles, New Mobility & the Built Environment Broadband Wireless

A Multi-Dimensional Broadband Network or Is This Just Pie in the-Sky?

[Added 5/17/20: The above video was modified from the original to include an additional animation of satellite-to-satellite communications via lasers, as described below.]

Could Starlink’s launch later this year signal the creation of a new Internet?¹ But, before we get there, the limiting factor for Starlink and its competitors might be the cost of the “pizza-box” sized terminal to communicate to the satellite constellation, according to CableLabs’ Principal Strategist (5G/Wireless), Shahed Mazumder.²  Could Elon Musk’s other company, Tesla, be part of the solution to drive down the cost of these terminals and help create a new wireless network that serves both rural and urban areas?

The Truly Mobile Radio #

That is, could Tesla integrate Starlink satellite terminals into their vehicles, effectively reducing the cost of the equipment, as compared to stand-alone terminals? It is conceivable that the antennas could be integrated into the body of the vehicle (pointed to the heavens, of course). The radio electronics would have to be added to the vehicles, but some of the signal processing might be performed with existing vehicle computational power.

Two companies that are taking a direct-to-existing device approach are AST, which is mass-producing satellites for launch from airplanes, and Alphabet’s Loon, a balloon-based system [2/5/21 -update Loon began winding down operations as of 1/21/21]. Both are working with existing wireless providers to create untethered cellular towers using frequencies used by existing handsets. They both have announced significant investments and/or partnership agreements with wireless carriers.

With a direct Starlink connection, Tesla would also no longer need telecom carriers for connectivity to their vehicles. This vertical integration would represent cost savings for the Palo Alto-based (at least for now) energy, mobility, (insurance and telecom?) company (see this interview with Princeton’s Dr. Kornhauser).

Real-time connectivity appears to be important to Tesla as a job page on their web site indicates that real-time connectivity is part of their Autopilot feature,

“Our networks learn from the most complicated and diverse scenarios in the world, iteratively sourced from our fleet of nearly 1M vehicles in real time.”

From a revenue perspective, this would allow Tesla to add another tier of broadband service to its existing offerings (which are basic and $9.95 per month). That is, they could conceivably offer a price-competitive home/mobile Internet service in rural and urban areas, to everyone who purchases a new Tesla product.

Added 6/9/21 – In March, SpaceX Services requested FCC permission to extend their service to “Earth Stations in Motion” meaning they will be mounted on cars, trucks, aircraft, and watercraft in the U.S. and around the world. “Consistent with SpaceX’s space station authorization, these ESIMs will transmit in the 14.0-14.5 GHz band and receive in the 10.7-12.7 GHz band.”

Where it gets interesting is the earthbound vehicles would serve as part of a land-based mesh network. In this scenario, the vehicles would double as base stations communicating both to other vehicles, in a peer-to-peer configuration, as well as directly to the satellite network. They would still serve as gateways to consumer devices. Of course, there is no reason these radios could not be included as part of a Tesla solar installation (e.g. a pizza-box antenna located next to the solar array).³

Added 5/17/20: In addition to distributed energy storage, perhaps distributed data storage is part of Elon Musk’s long-term plan. Each radio terminal could conceivably serve as a storage node in a peer-to-peer network. Perhaps Tesla charging stations will someday be both repositories of electricity and electronic files.

Added 6/19/24Elon Musk in his shareholder presentation suggests that 100 GW of compute capability will be available in the network of cars when it gets to 100M cars. He suggests even at 50% utilization this amounts to a lot of excess compute power that could be used to sell AWS-like services (e.g., the distributed data center envisioned in the previous paragraph).

Additionally, the low-latency promised by free-space optic transmission between satellites could be a major boon for international stock and commodity traders. Although it is reported that intra-satellite laser communications are not part of the initial launches a satellite to ground-based relay approach would still outperform terrestrial networks. University College London Professor Mark Handley simulated a link between Seattle and New York and found it would take 36 ms for an LEO satellite network versus 78 ms for a terrestrial fiber solution.

Lastly, it has been suggested by some that LEO networks could serve as alternative backhaul for existing rural ISPs.

Another Communications Plane #

Could this lead to a future layer, which would consist of Starlink receivers onboard air taxis? The above video speculates that Tesla could conceivably supply the air taxis (although that is highly doubtful, given Elon Musk’s view of the efficiency, noise, and the perceived safety of that form of transportation).

Added 3/9/21: SpaceX/Starlink is officially looking to bring its version of the Internet to aircraft, ships, trucks, RVs. Elon Musks says it isn’t ready yet for cars, but, could these other vehicles serve as mobile base stations?

Starlink could provide Internet connectivity to the many companies that aim to provide low-cost and ubiquitous regional air transportation.  Reliable and low-cost Internet connectivity for the passengers is probably going to be a baseline feature for air taxi services and this could be a sizable new market for a company like Starlink.

In turn, these Starlink-equipped aircraft could serve as relays both to the satellites as well as to earthbound vehicles. Ridiculous as that may sound, the Air Force is looking at such an approach for ad-hoc wireless networks in conflict zones. Starlink already has a $28 million contract with the U.S. Air Force Research Laboratory to test space-to-aircraft communications. And late last year Starlink demonstrated data throughput of 610 Mbs to a C-12 twin-engine turboprop aircraft.

The above video is pie in the sky speculation about the path Starlink could follow. What is clear is that, with 422 satellites already in the sky, they are well on their way to providing alternative broadband service beginning in 2020.

Footnotes #

¹Technically speaking Starlink would be an intranet, but from a user standpoint, it would feel like the Internet. In addition to the satellite to ground transmission via Ku (12 to 18 GHz), Ka (26.5 to 40 GHz), and V (40 to 75 GHz) bands, Starlink will use laser links via free-space optics creating a mesh network between satellites. Starlink is reportedly using a peer-to-peer protocol that, according to an Elon Musk Tweet, “will be simpler than IPv6 and have tiny packet overhead.” Implicit in this statement is that the Starlink radios will also serve as a bridge between its Intranet and the IPv4 and IPv6 protocols of the World Wide Web.

² In a February 2020 interview at the NCTC WEC conference, Mazumder suggested the cost of the terminals could be between $2,000 to $5,000 per unit at launch. Before declaring bankruptcy in late March. OneWeb, a SpaceX competitor, indicated to SpaceNews that terminals delivering community WiFi would cost between $1,000 to $1,500. In that same article, OneWeb suggested an aspirational goal of $150,000 for an airborne radio terminal, which they estimated to be half the cost of current airborne Internet radios [7/4/21 Update – OneWeb was rescued from bankruptcy and now has 254 satellites in the sky and is planning ISP backhaul services].

Musk suggested in a 2015 speech that the volume cost of the SpaceX/Starlink terminals will ultimately be between $100 to $300.

³ It is assumed that existing Tesla vehicles would need retrofitting while future models would have the antennas and radio hardware included as part of the design, Further, a land-based, peer-to-peer network would be similar in concept to Facebook’s Terragraph initiative.

Author Ken Pyle, Managing Editor

By Ken Pyle, Managing Editor

Ken Pyle is Marketing Director for the Broadband Forum. The mission of this 25+-year-old non-profit “is to unlock the potential for new markets and profitable revenue growth by leveraging new technologies and standards in the home, intelligent small business, and multi-user infrastructure of the broadband network.”

He is also co-founder of Viodi, LLC and Managing Editor of the Viodi View, a publication focused on the rural broadband ecosystem, autonomous vehicles, and electric aviation. He has edited and produced numerous multimedia projects for NTCA, US Telecom and Viodi. Pyle is the producer of Viodi’s Local Content Workshop, the Video Production Crash Course at NAB, as well as ViodiTV. He has been intimately involved in Viodi’s consulting projects and has created processes for clients to use for their PPV and VOD operations, as well authored reports on the independent telco market.

Linked In Profile

21 replies on “A Multi-Dimensional Broadband Network or Is This Just Pie in the-Sky?”

Ars Technia has an excellent summary of the FCC’s proposed order regarding the Rural Digital Opportunity Fund and how it excludes Starlink from applying for RDOF (Rural Digital Opportunity Fund) using the Gigabit performance and low-latency tier. Starlink can still apply for RDOF funds, which are allocated on a census block basis, but would be penalized, relative to solutions that deliver 1 Gbs/<100 ms latency round trip solutions.

Specifically, on page 36 of the order it says,

“The Auction Application System will not allow an applicant that intends to use any form of satellite technology, whether geostationary, high earth orbit, medium earth orbit, or low earth orbit, to select the Gigabit performance tier or to select low latency.”

In a footnote on page 37, it acknowledges SpaceX/Starlink’s testing with the Air Force (reference in the above article), but it reinforces its skepticism given that it isn’t yet commercial and widely deployed

“Given this lack of a track record of providing service to a mass market at any speed, no service provider that intends to use non-geostationary orbit satellites to meet its Auction 904 public interest obligations will be permitted to bid in the Gigabit performance tier.”

Similarly, on page 40, they question whether LEOS can meet the 100ms round trip requirements, again, given that it isn’t proven in real-world conditions

“In the absence of a real world example of a non-geostationary orbit satellite network offering mass market fixed service to residential consumers that is able to meet our 100 ms round trip latency requirements, Commission staff could not conclude that such an applicant is reasonably capable of meeting the Commission’s low latency requirements, and so we foreclose such applications.”

SpaceX/Starlink’s effort to partake in RDOF is reminiscent of Tesla’s use of tax credits (particularly trading of tax credits with other automobile makers) to provide critical financing in its early days. Could RDOF be a similar ploy by Starlink to effectively provide the financing it needs to get to mass deployment?

The FCC at its 6/9 meeting is allowing LEOS providers to bid using the low-latency, gigabit tier (page 26, section 73) They will have to prove it will scale, however.

“Specifically, we expand the network performance question applicable to fixed broadband wireless access networks to include satellite providers, and modify the question to require applicants to describe how the planned customer premises equipment will yield sufficient capacity. And we provide examples for the types of information we will be looking for in response to questions regarding base station configurations and channel bandwidths, as well as traffic and propagation assumptions. Moreover, we require applicants that propose to use satellite technologies to describe how the proposed network will achieve the performance tier(s) and latency requirements to all planned locations in a mass-market consumer service.”

And, as a reminder that there will be an ongoing operational cost, Rocket Rundown reports that two Starklink satellites have been deorbited for a failure rate of 1.2%. Whether this failure rate will be the long-term rate and whether that will ruin the economics of Starlink are questions that remain to be answered.

Here is Starlink’s latest filing with the FCC asking for Eligible Telecommunications Carrier status so they can offer lifeline service.
They need ETC status so they can receive the $885.5M RDOF funding to serve 643k locations ($1,134/location). As reported in another Viodi View article, many of these locations are in urban areas where there will be no reason for potential customers to take service (although Starlink will still get a subsidy for each of those locations).

They claim their 10k+ users are getting 100/20 Mb/s (downstream/upstream) service from their existing 1,000 satellite constellation. They don’t provide evidence, such as speed test reports, that their customers are seeing those speeds. They don’t explicitly state their long-term oversubscription ratio (which appears to be roughly 10:1, today). Assuming that they have 12k satellites at full-buildout (as referenced in the above Reuter’s link), and 100% RDOF subscribers (643k), that would mean a 55:1 oversubscription.

Starlink should be submitting data based on that sort of oversubscription ratio, not their current 10:1.

It is clear from their filing that some of the RDOF funding would go to launching new satellites:

“Significantly accelerate production of satellites and user equipment to ramp up capacity
deployed and CPE available. Accelerated production means more people receive service
faster in the unserved areas where SpaceX placed winning bids.”

Further, bullet four suggests they would be using the monies to buildout their telephony expertise.

“Invest in standalone voice service capabilities.”

It is assumed that Lifeline telephony operators have a voice network. Starlink shouldn’t be receiving RDOF funding for something that is a given.

I think Tesla with a Starlink interface is pie in the sky. Even though Elon Musk owns SpaceX (parent of Starlink) and is the major stockholder in Tesla, he is unlikely to extend Starlink access to moving vehicles. It would be too expensive and a bulky receiver attached to the vehicle.

I also do not think 5G interfaces will be built into any vehicle for the next 5 years or longer.

Could be too expensive for cars in the short-term, but there is definitely potential for delivery to aircraft, which is a market that could really take-off if the low-cost electric aviation takes off like many pundits think.

Starlink’s filing to the FCC for a small portable dish, as well as the earlier February filing regarding portability in cars, RVs, and planes, points to the idea that the Starlink receivers could be destined for what could end up being mobile cell towers in a P2P mesh, as posited above. With now 5 million vehicles world wide and the goal to get to a production of 20 million per year, this could be a real possibility.

And Amazon, with its Project Kuiper, has a good overview of its optical connected Low Earth Orbit satellite network. They are claiming speeds of up to 100 Gbps at up to 621 mile spacing. Like, Starlink, their plan is to create a mesh network that wraps the earth in connectivity with latency equal to 30% faster than the equivalent distance terrestrial network (it takes advantage of not having to deal with the earth curvature). They expect enough satellites to launch early customer pilots in the second half of 2024.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.