Ask HN: Why are flight blackbox data still not uploaded to cloud?
5 comments
Forget detailed blackbox telemetry, perhaps a more pertinent question to ask is why aircraft location data is not transmitted at decent frequency (60s?) via sat comms in a manner that is not able to be disabled via the aircrafts controls.
see: https://en.wikipedia.org/wiki/Analysis_of_Malaysia_Airlines_...
and:
https://en.wikipedia.org/wiki/Malaysia_Airlines_Flight_370#S...
Also for the other respondent about internet access, direct access is not required in many places as there are already extensive radio ground stations in use for aircraft telemetry.
Also for the other respondent about internet access, direct access is not required in many places as there are already extensive radio ground stations in use for aircraft telemetry.
Satellite tracking now exists: https://aireon.com/ https://www.radarbox24.com/space-based-ads-b
The bandwidth bottleneck is not on the cloud side; it's on the plane side. Satellite Internet from moving planes is quite expensive and would require retrofitting new antennas in many planes.
Many safety-critical systems do not use encryption because it's one more component that can fail.
Many safety-critical systems do not use encryption because it's one more component that can fail.
> Many safety-critical systems do not use encryption because it's one more component that can fail.
This is a good point but isn't some data still better than no data at all? If we decide not to collect encrypted data because encryption can break, then we have exactly no data if we lose the blackbox.
This is a good point but isn't some data still better than no data at all? If we decide not to collect encrypted data because encryption can break, then we have exactly no data if we lose the blackbox.
This is something that might well happen in the future. But to retrofit this to all commercial planes is not a simple or cheap task. Many planes in the sky were designed and built many years ago and so updating them to also upload to the cloud during flight requires substantial design and certification efforts that Boeing/Airbus are not going to suddenly take on unless they are forced.
Adding it as a feature of any new plane designed in the future seems simpler. Although uploading everything in real time be ideal you could easily ping with the essentials for almost no bandwidth. Every minute with location, speed, altitude and half a dozen other important numbers would be a good start.
Adding it as a feature of any new plane designed in the future seems simpler. Although uploading everything in real time be ideal you could easily ping with the essentials for almost no bandwidth. Every minute with location, speed, altitude and half a dozen other important numbers would be a good start.
If there was a war, it would be easy to shoot down all the planes.
What I wonder is - why not multiple black boxes - some at the end of the wings that break off, some that float off in water, some with dye packs, etc
What I wonder is - why not multiple black boxes - some at the end of the wings that break off, some that float off in water, some with dye packs, etc
Presumably there is a lot of data to be recorded and there is not consistent access to the Internet during flight? Seems like a fun problem to solve.
After doing some research into this matter I was surprised to find that we're still relying on oldschool physical blackbox to get flight data. But why? In this [1] Quora question a person claims to have gotten an answer from NTSB. They have two main arguments:
1. Data is too big that we don't have technology to deal with that much information
2. There are security concerns since people can intercept data.
I don't understand either of the arguments. Here are my thoughts:
1. Some basic search [2] leads us that there are approx in the order of 10k airplanes in the sky. Assuming there are no periods of time where flight traffic is orders of magnitude worse (since countries live in different timezones) seems like the problem we're trying to solve is streaming data from 10k clients. Assuming the biggest bandwidth bottleneck will be audio (since there is no video data and physical data is relatively cheap) we can reduce this problem to streaming movie from 10k clients at any given time. It seems like one way to verify this problem is tractable is counting how many people uploading video to youtube any given moment and finding out the delta between total video length and transmission latency. While I don't have any real numbers, it seems like a tractable problem.
2. Similarly, this seems like a problem that we already should be able to solve. AES encrypted keyboards, encrypted mouses etc come to mind; or even basic HTTPS.
Could someone please explain this layman why this problem is so hard? If someone were to attack this problem and maybe found a startup, what resources should they refer to?
[1] https://www.quora.com/Why-do-airplanes-still-use-black-boxes-while-all-data-could-be-saved-real-time-to-servers-on-land-through-the-Internet
[2]: https://www.quora.com/How-many-airplanes-are-in-flight-on-average-at-any-given-time-worldwide