Hacker News Reader

Comment by spuz | Hacker News Reader

The top comment from Arstechnica is:

> This will be absolutely fantastic for developing nations. Being able to bypass fiber-optic backbones for consumer connections will be a huge benefit. Imagine how many rural villages will be able to set up a couple of receivers to share among the residents.

Am I right in understanding that the intention is that SpaceX would act as the ISP in this satellite network? What are the implications for places like China where internet access is very much controlled by the government rather than some American billionaire? Are SpaceX going to breach the great firewall?

View on HN

Elon Musk has said that he doesn’t intend to fight governments over internet access, especially governments that can shoot down satellites:

"Obviously, any given country can say it's illegal to have a ground link... And from our standpoint we could conceivably continue to broadcast, I mean, I'm hopeful that we can structure agreements with various countries to allow communication with their citizens, but it is on a country-by-country basis. If they get upset with us, they can blow our satellites up, which wouldn't be good. China can do that. So probably we shouldn't broadcast there."

From: http://shitelonsays.com/transcript/spacex-seattle-2015-2015-...

Something people who have not worked a lot outside of the USA may not realize is: It's not hard for an autocratic regime to crack down on some types of WAN connection. Using currently available geostationary two way satellite technology, there are plenty of countries in the world where if you import and try to build your own Tx/Rx 3.8 meter size, C-band earth station and build a link to a teleport in Germany or France, the government will have a very simple response. A pickup truck with four men armed with guns will come and dismantle your dish.

I don't think this will be very helpful for scenarios like China at all. Or places like Iran where all the ISPs are forced to be downstream of the government ASN.

This will be very helpful for places on the planet that are extremely costly to reach with terrestrial fiber ($2m to $30m build cost) or a hill-to-hill relay of high capacity mountaintop tower sites with PTP microwave ($500k to $3m build cost to get from a major city into the hinterlands of some african nations). Also for places like small pacific islands that can never hope to economically bring a submarine fiber optic cable to them. A lot of these places have already moved off "traditional" geostationary C and Ku band transponder leases and onto o3b. This is basically a similar concept to o3b but applied not only to large backbone links for ISPs, but to individual customer premises rooftop antennas that might have a cost of under $1000.

>> there are plenty of countries in the world

citation please? Do they take it down because you don't have a license to transmit in that spectrum band?

Perhaps (see sibling comments), but I don't think that's what GP was getting at. They take it down because the government (or whoever is in control) wants to control the flow of information amongst their citizens and not some private (likely foreign) company. What good is the Great Firewall if you can just subscribe to Elon's ISP with a dish?

Feasibly, SpaceX would not need to negotiate or conform to any nation-states censorship requirements, because the citizens can just directly access the satellite through private infrastructure.

A capitalist analogy is how hotels used to (and many still do) charge outrageously for internet. When people started using their own infrastructure to connect their laptops to the internet, hotels (like the Marriot[0]) jammed wifi hotspots.

[0] http://edition.cnn.com/2014/10/03/travel/marriott-fcc-wi-fi-...

Sudan, Iran, China, North Korea, Pakistan, Uzbekistan, tajikistan, Turkmenistan, Kazakhstan, Venezuela just off the top of my head.

> Are SpaceX going to breach the great firewall?

I don't know much of anything about international law, but I don't think that would ever work. Countries have sovereignty over their airwaves. If they require a permit to broadcast you need a permit to broadcast, even if the signal is coming from space. Countries have sovereignty over business conducted in their territory. China can just say that SpaceX's internet service is illegal. How should SpaceX get their illegal money out of China? What to they do with the (still illegal) money once it is out of China? How does SpaceX get ground stations for an illegal internet service into China?

Maybe SpaceX decides to give away free internet in China as a humanitarian publicity stunt, and flaunt the Chinese Government. They could publish the design for the base station and hope that sophisticated dissidents manufacture them. Now China uses every diplomatic avenue available to try to stop this: the UN, the ITU, espionage, treaties, whatever. Is the US going to cover for SpaceX? Will the state department decide to burn all of their bridges to the largest country on earth for the benefit of a publicity stunt? If China shot down SpaceX satellites, would they be justified in the eyes of the international community?

SpaceX probably could try to do this, but would be faced with a few challenges. As a two way satellite engineer, just off the top of my head:

a) smuggling/importing the rooftop terminal equipment into china

b) if not smuggling, finding a way to use greymarket resources to manufacture the terminals in China

c) making it small enough and easy enough to hide that it can't be obviously spotted/removed from a roof. Google a bit about Iran's periodic and mostly unsuccessful forcible removals of Ku-band TVRO (Rx only) TV satellite dishes for an example. China has a lot more manpower and police-state type resources to deploy to hunt down antennas. If you are faced with the problem of requiring clear unobstructed line of sight to as much of the sky as possible, while also hiding it, you have a technical problem.

d) It is not hard for a police state like China to train relatively unskilled enforcers to use directional spectrum analyzers to hunt down unauthorized rooftop Tx, anywhere from 6 GHz to 80 GHz.

Thanks, this is the kind of insight I was hoping for. You answered lots of questions I wouldn't think to ask.

> Countries have sovereignty over their airwaves.

Is this true? Isn't, say, a San Diego FM station violating the sovereignty of Mexico then?

I would hazard a guess that there's treaties for this sort of thing.

The US has treaties with Canada and Mexico for this. Both for AM/FM radio and for things like FCC part 47 and part 101 licensed microwave links.

China could just legislate that the service be re-sold by a local ISP with all the required encumbrances.

I'm get to be convinced there is a technical solution to the problem(s)(?) of government / legislation / regulation / right holders / etc etc.

Progress (movement, anyway) happens over time with push and pull from both directions / all stakeholders.

It would be pretty hard (from the technical and legal perspective of a drinks-the-koolaid network engineer at the Chinese version of the FCC) to build a fully compliant, tamperproof version of the GFW into a standalone CPE rooftop antenna+indoor router combo. The Chinese GFW architecture relies on the government having pretty strict control over the major ISPs (top 20 to 30 AS numbers in the country by size of APNIC v4 and v6 blocks, number of peers, etc) that have capacity in the submarine cables linking china to other southeast asian nations, and directly to the USA. Also on the terrestrial cross border fiber.

> fully compliant, tamperproof

Of course, but the idea is to simply make the only legal and easily accessible version regulation compliant, and then that is probably what most people will use.

Satellite Internet exists already. There are no new implications for SpaceX's implementation in China that don't apply to the incumbents.

I'm not sure that's true. The big player, Hughesnet, has satellites in geosynchronous orbit over the US. You can't "see" those from China.

SpaceX's plan is unique in that they aren't planning to run satellites in fixed, geosync orbit, to my knowledge nobody is doing that yet.

No, it is true. There are many, many satellite internet (VSAT) network operators. These are companies that lease bandwidth from the satellite owners and operate a network within the confines of the lease. The concept that is being discussed is landing rights, where a satellite operator is not allowed to import or sell their products in a country. There is also a matter of coordination, which means that countries with satellite systems that are higher up in the first-come-first-serve international regulatory list can object to those that come after them.

From a regulatory perspective, SpaceX is not not doing anything that Iridium, GlobalStar, Inmarsat, Intelsat, SES, etc...have not already dealt with.

No, from a regulatory perspective and technically this is a very different thing. Aiming at LEO and MEO satellites that move across the sky is very different than aiming a fixed-mount VSAT dish at a geostationary location. I've built large C and Ku band earth stations and worked with some of the largest teleports on the planet. This is something new and novel.

The SpaceX concept intends to use phased array/software beam steered antenna systems for the individual end user CPEs.

If you're lumping Globalstar (a mostly epic failure of LEO bent pipe low bitrate satellite constellation) in with SES (a company that operates $200m, 4000 to 6000 kilogram sized geostationary C/Ku/Ka band satellites and is highly successful worldwide), you're grossly misinformed, they are totally different architectures.

Technical architecture is independent of regulatory frameworks. The ITU simply sees this system as a non-geostationary system. The only alternative is geostationary. The parent was connecting on novel regulations which would need to be enacted, but new regulations are not necessary as existing ones cover this new system just as they do old ones.

geostationary satellite already exists. This is a very different thing.

MEO by o3b already exists, has been running for a few years, but is only available for larger telecoms and ISPs (example: a telecom in a pacific island nation state that doesn't have any submarine fiber). It's a great technology but the minimum size terminal is a pair of motorized, tracking 1.8m size antennas. It's an alternative to multi-ten-thousand-dollar-per-month C and Ku band transponder kHz and MHz.

Nothing like this already exists.

And how great is it going to be for the developing parts of the US!

Reply on news.ycombinator.com