In-flight broadband’s story so far has been similar to that of airplanes sitting on the runway, waiting for clearance to take flight. Despite a big push from Boeing and other major international carriers, in-flight broadband was stuck on ground, burning dollars like an idle plane burns gas. No more! Thanks to new surface-to-air technologies used by companies such as Aircell, the business is ready for takeoff.
Of course, airlines’ desperation to make money any which way they can has helped accelerate the rollout of GoGo-like services. In-Stat, a market research firm, predicts there will be 800 planes with in-flight broadband by the end of 2009 vs. just 25 in 2008, generating $47 million in global revenue. By 2012, In-Stat expects in-flight broadband will be a billion-dollar-a-year business, with demand for in-flight broadband equipment nearly doubling between 2009 and 2013. Indeed, airlines such as Virgin, American Airlines (s amr), Alaska (s alk) and Delta (s dal) are being super aggressive with their rollouts.
By 2013, In-Stat sees the number of annual in-flight broadband connects topping 200 million. “Connections from handheld devices will account for about 1/3 of connects, with notebook computers accounting for 2/3,” the firms predicts. On that point, I think they’re wrong. It will be the other way around. Using the iPhone/iPod Touch to connect and get emails is a much better and easier option than opening up a notebook.
Useful link: The GigaOM In-Flight Broadband Cheat Sheet.
11 thoughts on “After Long Delays, In-Flight Broadband Is Taking Off”
i used the service recently on a virgin flight.
they have a good iphone interface that makes it easy to get online, and the network is pretty fast too.
was able to browse webpages, get email, and even saw a youtube video !
kudos to aircell
How much total bandwidth do they have to the plane? No one ever answers that question. That’s the gating factor here because it’s the shared resource.
If you’re sharing that connection with 100 other people (who are actually using it) they better have a pretty fat pipe for you all to share. Early experiences where there are only a few passengers online are pretty moot as far as speed is concerned.
Erik, not sure who “no one” is. I report on in-flight broadband for a variety of publications and my own site, and I’ve interviewed the execs at Aircell, Row 44, and other companies many times. Whenever I ask them about bandwidth, they tell me. Then I publish that information. I’ve seen the numbers in a lot of places.
The marketing folks (and legal folks), however, don’t like to put that data on an in-flight info card, or tell you when you call, because then it’s a promise, and it’s hard to ensure they deliver it.
Aircell is currently using EVDO Rev. A set to its special 800 MHz frequency allocation they won at auction. EVDO Rev. A has a top speed of something like 3+ Mbps in terrestrial uses with 1.25 MHz of spectrum allocated. Aircell won 3 MHz of spectrum (1.5 up and 1.5 down), which allows them to use Rev. A in that framework. The company has often said they can get a couple Mbps to the planes.
In reality, how much can that be with lots of users? The issue will be how they shape connections, and how many people are trying to use YouTube or download videos, versus more traditional activities. A lot of studies show that people adapt their behavior to the pipe that’s available. So once you discover YouTube isn’t working or whatever, you stop doing that and switch to another activity.
I was on the launch PR flight for Virgin American in November, and there were about 130 people on board all of whom fired up 1 or more Wi-Fi devices. The network performed quite well under that crazy load; how, I don’t know. I was able to see 700 Kbps downstream in the middle of the flight, even as I saw people doing all kinds of high-bandwidth activities. So Aircell’s packetshaping must be quite good.
Row 44 is claiming about 4 Mbps of TCP data and a larger amount of streaming video capability. But that’s yet to be proven in flight.
So you’re talking about in the area of 5 Mbps tops.
That can support loads of people doing email and web browsing.
Hulu, not so much.
Something where latency matters like Skype? Seems even less likely.
Another question; There are huge swaths of the American west that are not covered by EVDO (Nevada, Utah, Wyoming, etc, away from the cities and the highways), does it just stop working there?
This is not the same EVDO network provided by Sprint and Verizon, it is a completely separate network(s) operated by these in-flight carriers. Separate frequencies. Separate antennae. So, just because an area doesn’t have EVDO on the ground doesn’t mean it won’t have EVDO in the air.
What Jesse said — EVDO Rev. A is the technology, but Aircell has deployed its own network with the antennas all pointing up at planes. It’s a different frequency than any cell network is licensed to use, too; an exclusive air-to-ground allotment.
Aircell is already talking longingly about LTE, which will let them dramatically boost per-plane capacity.
For services based on satellites, the downstream capacity to a plane can approach 80Mb/s. However, it’s unlikely you would have 100 people on a plane using the service. Despite it being perfectly suited and highly attractive to the demographic represented by people reading this blog, the usage numbers is lower than you might think. It will grow, for sure, but expect the active users on a 130 person plane to be under 25.
The 80 Mbps isn’t being used Row 44 any more. The last time I talked to the head of the company, he had scaled down that number to 4 Mbps for Internet and something higher for video streaming. 80 Mbps is a theoretical top rate that includes overhead and doesn’t include error correction, and other factors.
Also, with Ku-band (geostationary orbit) satellites, each transponder is pointing at a sector. If multiple planes are in the same sector, which is likely, then bandwidth to each plane is shared unless more transponders are used.
Glenn you might want to note that the latencies getting from and to geo orbits will also gargle balls.
In March of 2009 I was very excited to take my first GoGoInFlight trip on Virgin America. However, among other accessibility issues, I found a visual only captcha. A captcha of course is a displayed image where the user must enter characters they see in the image in order to authenticate. In this case, captcha is required in order to connect to the internet with GoGoInFlight.
The problem with all this is that I am blind, and screen readers do not read visual only captchas. You can view a video which demonstrates this here:
I contacted the company, and things were (typically) slow at first. They wanted to essentially thank me for my suggestion, and send me on my way. However, eventually, I was able to get in contact with someone other than the Twitter representative or the customer care department.
I received voicemails that, now, the company would locate a business owner to look in to this serious accessibility issue.
I expected, any day, to receive a phone call or email that the company had an idea about either a short term or long term fix for the captcha accessibility issue. These solutions might include, for example, audio captcha, using a different question rather than an image, or setting the system to bypass the captcha requirement for users who had requested this accommodation.
Instead on Friday May 1st I received a profoundly disappointing email from a director in the company.
First, he trotted out words like “safety” and “security,” to justify the cpatcha. He insinuated that some type of law, regulation, or other ordinance may require captcha. As it turns out, a competing service called Row 44, which provides wifi for Southwest Airlines, does not have a cpatcha.
Accesibility aside, it is baffling why goGoInFlight feels the need to have a captcha at all. Do they think that hackers will try to break in to the system so they can pay for the service?
Anyway, the email continued that the company planned to place alt-text on the submit button for the captcha form saying, basically, that if you are blind you should ask a passenger or flight attendant to help you with the captcha.
In a response, I articulated to GoGoInFlight that this would be akin to a restaurant opening with steps, and erecting a sign saying, basically, if you are in a wheelchair, ask someone to carry you up and down the steps! Whether you agree with it or not, American disability civil rights laws have a higher standard of accommodation required. A cafe cannot just open, and tell people in wheelchairs to get someone to carry you up the steps.
Polling places cannot just tell the blind to get someone to help you vote. The HAVA law requires, in fact, that accessible voting machines be deployed.
From a purely technical perspective, there are many mobile devices that blind people use that do not have a screen. I don’t know how GoGoInFlight thinks I can get someone to read me the image on a device without a screen. I had such a device with me on my March flight:
I serve as the president of the National Federation of the Blind of Massachusetts, and I have made this GoGoInFlight issue a top priority of my administration. We will not rest until blind people can access the internet on these airplanes free of barriers like captcha.
We are engaged in a number of public relations, administrative complaint, media, and other activities in order to bring this issue to resolution.
Was on luthansa couple of years ago, which 1st launched inflight broadband. passenger nearby was skyping. plane drone makes one talk LOUD. Air rage wage was percolating in that cabin. anyway, OM is right – it’ll be more like 2/3 users on smartphones vs laptops/netbooks….