Analyzing Flightradar24’s internal API structure. | Hacker News Reader

Analyzing Flightradar24’s internal API structure.

blog.cykey.ca

36 pointsCykey12 comments

I would love to see an open source network of ADS-B collectors form, similar to what's available for the APRS-IS [1] network. The big ADS-B players (Flightaware and FlightRadar24) get their ADS-B feeds from the FAA (5-min delayed) and supplement it with real-time over-the-air collection from volunteers around the world. These companies give away small ethernet-connected collection servers to the volunteers and in exchange, they get the data and keep it proprietary.

It sure would be nice if there was a network to collect and pass around this data for free consumption by whoever wanted to use it.

[1] http://aprs2.net/ http://www.aprs-is.net/

Indeed, it would be nice. Something like that shouldn't be too difficult to do. People all around the world would send their ASD-B data to a centralized server which would process it all and make it available via a nice, stable API. Something like http://www.openstreetmap.org/.

ADSB# has the ability to stream data to sdrsharp.com, but I don't know what they do with it.

ADS-B won't be awesome in the US for another 5 years, though. My office is at one of the busiest regional airports in the country. Very rarely do I see any aircraft that provide more identifier and altitude using my RTL-SDR and homemade antenna with ADSB#.

Your antenna is probably bad. I live in the middle of nowhere USA and get planes all the time. (Using rtl_adsb.)

Regardless, his statement is true that ADS-B is not sufficient for full coverage in the US right now. I've seen estimates as low as 25% for the proportion of US carrier aircraft currently ADS-B. FlightAware in particular uses far more than just ADS-B.

Can you share how many and how far away? Are you near an ADS-B ground station (which will rebroadcast all traffic if a single ADS-B squitter/output is near) or only receiving active output from airplanes?

The ones I do see are typically freight carriers (FedEx, etc), or international airlines (Virgin). AFAIK, most of the flights into and out of this airport aren't those types. So, of the air traffic I do see, the majority don't identify with more than identifier and altitude. Lat lon is few and far between.

Very true if 5 years even makes it popular. Most regional carriers and even legacy carriers do not have ADS-B out in them. The adoption rate in general aviation is dismal. I believe under 5% is what AOPA is saying, my personal experience is 1 guy out of 30 owners I know has installed it. The benefits just aren't there yet.

Flightaware I wasn't aware of, though PlaneFiner[1] is another good resource.

[1] www.planefinder.net

Just sent you an email about this...

I wrote a small wrapper javascript library around the planefinder.net API: https://www.npmjs.org/package/planefinder

I'd love to see a free & open network of shared ADS-B data, but the hobby seems to be stuck in a proprietary mode. For example, Planeplotter[1] runs its own proprietary network for customers only, which includes enhancements like multilateration-based localization of aircraft that are only using Mode C transponders, without GPS coordinates.

All the aircraft information databases that tie the ICAO hex codes sent in ADS-B data are proprietary, and there's often dumb drama over incompatible versions or updates.

[1] http://www.coaa.co.uk/planeplotter.htm

I'd love to see a post from Flightradar24 talking about their internal tech, about how they deal with updates and requests from around the world which only make sense in real-time. Maybe some series of message-queue based processing (or equivalent, like Storm?) publishing into something in-memory (like Redis?). Combined with the zones approach, means there's a minimal amount of processing on each request. Just a guess, though.

Reply on news.ycombinator.com