Satellite Internet? Hmmm...
The Market Ticker - Commentary on The Capital Markets
Logging in or registering will improve your experience here
Main Navigation
Display list of topics
Sarah's Resources You Should See
Sarah's Blog Buy Sarah's Pictures
Full-Text Search & Archives
Legal Disclaimer

The content on this site is provided without any warranty, express or implied. All opinions expressed on this site are those of the author and may contain errors or omissions.


The author may have a position in any company or security mentioned herein. Actions you undertake as a consequence of any analysis, opinion or advertisement on this site are your sole responsibility.

Market charts, when present, used with permission of TD Ameritrade/ThinkOrSwim Inc. Neither TD Ameritrade or ThinkOrSwim have reviewed, approved or disapproved any content herein.

The Market Ticker content may be sent unmodified to lawmakers via print or electronic means or excerpted online for non-commercial purposes provided full attribution is given and the original article source is linked to. Please contact Karl Denninger for reprint permission in other media, to republish full articles, or for any commercial use (which includes any site where advertising is displayed.)

Submissions or tips on matters of economic or political interest may be sent "over the transom" to The Editor at any time. To be considered for publication your submission must include full and correct contact information and be related to an economic or political matter of the day. All submissions become the property of The Market Ticker.

Considering sending spam? Read this first.

2019-05-16 12:05 by Karl Denninger
in Technology , 159 references Ignore this thread
Satellite Internet? Hmmm...
[Comments enabled]

There's a chuckle factor associated with all of this "breathless reporting" on new attempts to bring "satellite" internet to the world.

Upcoming satellite technology could lower prices for Internet services, a new report predicts.

Low Earth Orbit (LEO) satellites from Elon Musk’s SpaceX Starlink project and Jeff Bezos’ Project Kuiper could save American households more than $30 billion per year by introducing more broadband competition, according to a report from BroadbandNow.

“The arrival of this emergent technology is likely to drive down monthly internet prices for hundreds of millions of Americans,” the report said. In short, the more broadband Internet services available in an area, the lower the price consumers will pay on average.

Weeeelllllll, maybe.

Wayback machine time.

When I ran MCSNet I had a little "skunk" project that was my personal pet and nobody else had access to, nor even knew existed.  I detected quite early on that the technological progression for Internet access was likely to go from dial-up (where we were when we started for ordinary consumers) into the monopolist-owned pathways.   ISDN became available quite early on but was still 64kbps per channel, or 128kbps in your house.  Plus compression it wasn't bad and latency was much better than a modem, but still.... and it was expensive, both for the terminal equipment and for the circuit, which was charged like a phone call (or 2 in the case of 128k bonded.)

The "promised land" was sold by many as DSL, but every analysis that I did said that the companies selling it at the time (Covad, etc) were all going to go bankrupt.  They did, ultimately, and I was very, very happy we never got in bed with them as the hit to customers when it happened was extreme.

The risk was from the cable companies which could (illegally) cost-shift much of their build-out into the cable TV space and just like today with health care and online product sales (e.g. Spamazon) I saw zero evidence that state or federal AGs would go after these guys on a competition basis.  This was an existential threat down the line a decade or so to the industry -- and a threat that I fully expected to materialize.

So I spent a lot of time noodling on ways to beat the monopolists without shooting their CEOs or the State and local governments and "regulators" for their refusal to do their job since shooting people was at the time (and still is) illegal.

One of the concepts I played with was building an embedded-system box that you'd buy and plug into your home.  It would run on unlicensed RF spectrum and, when plugged in, look for other reachable nodes and insert itself into a mesh.  At the time (pre-IpV6, which was ratified as a Draft in 1998 -- there was no generally-usable stack available prior to roughly 2000) we had moderately-severe constraints on IPv4 address allocations, and there were business-level "wars" that broke out from time to time.  There were also moderately-severe and recurring routing table problems, although the most-serious of them was quite-early when the first AGS+ routers simply ran out of addressable RAM space, and being VME/M68k based products couldn't have more RAM added to them (they got forcibly turned into "mid-level" and edge routers, and ultimately were scrapped -- at very significant cost.)

My scheme to address this was to layer an internal, network-private prefix in front and carry it only in the aggregate for the specific metro area; it would not be visible to anyone either end of the jack, but was essentially an extension of address space for the purpose of local area aggregation and traffic management.  The design was intended to deliver auto-configured Internet service out a standard Ethernet jack on the back of the box.

The problem that killed the project was the lack of a reasonable cost-and-availability solution for the processing and RAM required in each of the nodes to handle what was a non-centrally-controlled (that is, self-configuring, self-healing and self-modifying, yet secure against tampering) mesh.  It wasn't an unsolvable problem per-se, but on a reasonable money per box basis it was, considering that the mesh itself was worthless without external connections, which would have to be charged of course at various "touch points."  I still have working code and my notes; I never filed a patent on it or disclosed exactly how it worked to anyone because my calculations were that by the time the horsepower necessary to make it actually work in the consumer world was available the patent protection time would have either run out or been very close to expiration, and thus there was no way to make money at it.  As I expected would happen the cable companies indeed came in, leveraged their existing rights-of-way and monopolist access to same plus their ability to cost-shift and here we are.

These new LEO schemes share similar characteristics, except that of course today cheap wide-scale CPU and RAM power is not all that tough of a problem, and now you have formal IpV6, so at least in theory addressing isn't a problem either.  Unlike my idea of blanketing an urban or suburban area these systems would not care about density of customers on the land side at all.  But what IS a problem in the LEO space is that since the satellites "move" from the perspective of the ground station you need a phased-array antenna with a fairly decent amount of processing power to manage it and you also need the spectrum which by its nature is concentrated at the satellite end instead of being distributed as is the case with a mesh.  And finally the satellite side has constraints too -- much of them power-related, as flying birds up there have fairly severe energy constraints; you can't exactly plug something into the wall, so you're limited by what you can harvest with solar cell arrays.

I'm not necessarily sold on the will be cheaper paradigm, even if Space-X manages to get the cost of lofting the birds down to the degree that Musk thinks he will.  The birds still cost a lot of money in capital expense to build and the bandplanning required for such a system along with the processing and protocol work so as to make the network functional and solve potential interference problems is not trivial.  It's not an unsolvable problem by any means, and the latency of a connection in LEO as opposed to working in geostationary space, never mind the slot limitation issues with geostationary orbits (at 1/2 degree spacing you only have 720 of them, etc) that evaporate with an LEO constellation do help a great deal, but in the end it all comes down to actual deliverable bandwidth .vs. cost.

One of the problems with the existing option -- HughesNet -- is that not only is there a latency issue due to geostationary satellites being used but in addition there are extremely severe uplink speed limits and a "fairness" algorithm that prevents the sort of use that people enjoy today on cable systems -- such as watching 4k Netflix shows.  Additionally, due to the inherent technology with geostationary birds and that you're running a licensed-band transmitter there is a very real potential issue with interference and as a result "professional installation is strongly recommended" (so to speak.)

For this sort of system to replace HughesNet those issues have to be resolved in a way that actually makes sense.  The service has to be cheaper, latency lower, the "fairness algorithm" either gone or greatly attenuated and real two-way (that is, I have to be able to get a reasonable uplink speed that I can actually use) service has to exist. These are not trivial problems.

However they're also not unsolvable problems, providing the financial side works, which I'm not sold on.

In any event were I someone with a financial stake in Hughes I'd be real worried about right now -- if any of these systems actually go online and work as the developers think they will those folks are done.  If two or more them get operating, and they're not connected in some business sort of way, then real competition might show up.  Maybe.

However, and this is a big however, I'm not sold on this being a better mousetrap all-in cost-wise than, for example, mixed-band 5g.  Much is going to depend on installed density; in a rural part of the nation, for example, it is not established which gives you a better all-in cost structure.  Installing terrestrial backlinks and transmitters of course only serves that specific area, which is where the LEO system appears to "win", but the density required for real service to work in an LEO constellation is quite large because by definition those satellites are not in fixed relationship to a space on the surface and thus they only work when the necessary density is present so that the earth and space sides all can "see" each other and be able to pass traffic on a reliable basis, handing off and even passing traffic between satellites as necessary before the relative motion of the antenna on earth and the one in space degrades the signal to the point that the user notices it.

Again -- this isn't a "mobile internet" sort of application that will talk to your phone, as the phased-array antenna requirement (we're talking antenna sizes of perhaps a foot or two on a side, plus the power to operate it) will prohibit it from being used to connect to mobile devices.  But it's certainly something that could trivially blanket rural and even very rugged areas where horizontal "lines of sight" for terrestrial RF is a problem with quality, low-latency service.

IF the numbers work -- which remains to be seen.

View with responses (opens new window)