Slashdot Mirror
Airbus, Delta, and Sprint Are on a Quest for In-Flight Wi-fi That Actually Works (fortune.com)
It's 2018, so why is it still seemingly impossible to get a decent wi-fi on an airplane? From a report: Well, a lot of reasons, it turns out. The Wall Street Journal recently enumerated them: hardware, software, government regulation, aviation regulation, and rivalries between wireless and satellite companies. Despite the obstacles, a new alliance between Airbus, Delta Air Lines, Sprint, and two U.S. satellite companies is trying to find a way to provide faster Internet and a better user experience. Japan's SoftBank, which owns 80% of Sprint, and India's Bharti Airtel are also reportedly supporting the project. The group, which calls itself Seamless Air Alliance, envisions a world where a variety of devices could easily connect to the Internet while in flight at industry-leading speeds, rivaling cable and 5G. The businesses that are either involved in or backing the alliance pack a punch: they already serve about 150 million airline passengers and 450 million mobile users around the globe.
It bounces from cell tower to cell tower, but it seems to work.
Probably like $10 for one hour of wifi. At least if they get it to work well with decent speeds it'll be worth it.
Me, I'll just use Elon's low earth orbit satellites. It should work from inside an airplane, right?
Someday you might be able to get internet satellite service. Would in fright WiFi matter at that point?
I'll see your senator, and I'll raise you two judges.
They have a proprietary system. Don't know how it works. But you get it from gate to gate and it's fast enough for WebEx meetings with no audio drop outs. (And before somebody comments, I don't talk on those meetings, only listen. If I have something to add, I use the chat feature)
We know that as soon as the aviation problem is solved, they'll remove airplane mode. Eventually, we'll have constantly connected devices with no way to turn off aside of maybe waiting for the power to run out (then again, wireless charging could prevent that from happening).
Cable and 5G? I'd take DSL quality speeds if it came with with reasonably low (and consistent!) latency. Say, 250ms to popular sites.
I get to tell the office that I'm unreachable for a few hours.
Dear passengers, this is your captain speaking. We are just crossing over into China so all Wi-Fi will now be disabled in-flight. Thank you, that is all.
one cable node to a plane with 1-5 users willing to pay = good speeds.
JetBlue has free in-flight wi-fi right now. Wikipedia says they're served by Exede, a satellite internet company which also powers internet on United and Virgin America. Fast enough to stream NetFlix (at least, they claim so, I've never tried) and definitely fast enough for my recreational browsing. I'm not sure why Delta and Airbus would be late to the game, but I recall reading somewhere that their current in-flight connectivity is via a network of ground-based stations (probably the source of the Sprint connection), so they may be stuck with a different web of contractual agreements and regulatory hurdles than satcom.
(JetBlue also has free snacks and drinks and decent legroom in coach, which is why they'll always get my business if they're flying where I need to go.)
Maybe throw in a $10 GPS chip streaming co-ordinates to a base server so you can find your multi-million dollar jet next time it goes off the radar.
Seriously. Why has this not been an FAA requirement already?
...but fairly useless for anyone who has ever tried to do any sort of work on a flight using anything but the smallest netbook. It's been years since I last found that there was enough space in an airline seat to even open up a laptop. Time on the plane is better spent sleeping or listening to music---neither of which need wifi.
CUR ALLOC 20195.....5804M
Wouldn't everyone just prefer ethernet for their laptops?
At one point, Airbus owned the Norwegian firm called Marlink. They are one of the world's largest satellite communications companies (also a customer of mine). These guys are not "newbs" and they reliably deliver internet connectivity with relatively decent bandwidth to ships and oil rigs around the world. They are certainly capable of delivering in-flight wifi as well.
As most satellites these types of companies use are generally of the passive types (signal reflection) that means that so long as the ground (and air) based transmitters stay within leased bands, they can change modulation pretty much however they choose to increase bitrate within the same spectrum.
Latency for geostationary satellites will have long latency... that's physics... so sorry. But for LEO, the main problem is availability of satellite networks. This of course will change considerably now that SpaceX is launching. Viasat seems to be competent in the area of in-flight internet using this LEO satellites, however new technologies should be substantially better.
The main problem I've seen with satellite coverage has be Fresnel related. People love to blame clouds and such, but since most satellites are generally launched and aimed for providing maximum coverage at sea-level (plus or minus a few thousand meters), there are often gigantic uncovered gaps in coverage as altitude increases... there is issue that satellites are aiming to cover earth's surface. Consider the shape of a satellite signal reflecting to earth to be somewhat cone shaped. Then consider enough cones aimed at the surface of the planet to cover the entire surface of earth with about 20% overlap at the surface. Then consider that LEO is typically 160km to 2000km from earth. If a plane is flying at about 12km, The overlap is obviously reduced if not leaving gaps between cones.
Picture this differently, take a ball a little less than 13km wide with bumps that protrude at most 8.9km, now blanket that ball with the absolute least number of satellites possible to achieve near 100% coverage by aiming cone shaped signals from the satellites at the surface of the ball. The closer you are to the ball, the lower the latency. The further from the ball, the higher the latency. If you get too close (less than 160km) from the ball, if satellite will burn to ashes at it falls to earth. If you get further than 2000km, you're no longer in LEO which has it's own problems. Even so, a distance of 2000km from earth has a latency substantially longer than 160km (and no it's proportional as the distances must consider more than just up and down).
Once you have surface coverage, you need to consider how you will communicate with all the satellites. On option is using multiple space ports around the world to aim at all the satellites broadly and have the reflect. That would require entire bands of licensed spectrum that could only be used by a single network in the world. If we ignore the technical issues with this and then try to handle just the regulatory issues, this will never be a good solution. You'd need to license the entire sky all around the world for the bands you wanted to spam.
So, the next issue is how to get coverage from satellite to satellite. This can be done using active relays of signals between satellites. For geostationary orbits, this would be REALLY REALLY REALLY slow and REALLY REALLY REALLY REALLY expensive to use satellite to satellite communication that is "active" in the sense that the satellite would be able to store and forward packets actively. This is because as radio technology advances, new radios would need to be installed on the satellite to handle it. On LEO satellites, it's possible because the launch cost is far less, but the number of satellites to produce the same coverage is substantially greater, so the cost benefits are probably lost.
I guess it's already past the point of documentary length now. In any case, let's just say the problem is not trivial and even after 50+ years, we're still noobs on the topic.