Hacker News new | past | comments | ask | show | jobs | submit login
Airframes.io an aircraft-related aggregator for ACARS, VDL, HFDL and SATCOM data (airframes.io)
146 points by lsllc on Jan 26, 2023 | hide | past | favorite | 39 comments



Ahoy! Thanks to whoever posted a link to Airframes here on HN. I'm the founder of the project.

It's been a bit of a crazy 24 hours, and I haven't had enough time to fully surface and understand the impact that the ADSB Exchange sale has on everyone, the community, and Airframes.

Most of the Airframes effort has been fairly heads down, built in a tiny corner of the community and was not prepared for the events of the last day. Until now, most people weren't even familiar with technologies such as ACARS or VDL, limited to a mostly a small cross section of the community.

It has been a careful and deliberate effort to grow feeders & ingests slowly due to the nature of the data and infrastructure/storage needed, from both discovery (figuring out the data structures) and process (how to make use of it) perspectives. Requiring much of the efforts on the backend that the currently simple web app does not exactly reveal.

A new web app in development (there are some obvious and glaring issues/quirks with it now), and several other components, such as a desktop app, a mobile app, and a multi-architecture radio-focused OS to easily setup feeders (to Airframes, and the other aggregators) that will expand to other radio interests in time. There has been a lot of preliminary work on each of these.

The plan is to open source much of this over time.

If there is interest, I'm happy to elaborate more. I have been very transparent about the development process and implementations on the Discord in realtime. You are welcome to explore the dev channels there to get more background in the meantime.

Note that due to current events, I am taking on higher traffic than usual, unexpectedly, and everyone is still trying to understand what the impact of the recent events are.


There's quite the chorus of "Everyone is moving to airframes.io" or other similarly phrased posts in the ADS-B Exchange buyout thread[1]. Based on what I'm seeing, I feel like this is probably a premature place to settle on.

1. If you go to https://github.com/airframesio/ , it talks about how they want to eventually open source stuff. But mercy, like 90% of their stuff is private right now.

2. The Airframes roadmap has basically not been touched since 2019. https://trello.com/b/1vq5pHNq/acars-server-public-board (linked to as the "Airframes Public Trello Board" on airframes.io about page, both owned/authored by Kevin).

3. I also have found no public API documentation for reading or querying data. There appears to be a webapp one could pretty easily reverse engineer, but I haven't found any ToS or usage documentation that grant any rights to use the data.

I would absolutely not go to airframes.io in it's current state.

Side note, I think this is a super fascinating use-case for what p2p might want to target. Trying to draw together & perhaps correlate & index many highly active streams of data is a very compelling (and hard) problem.

[1] https://www.jetnet.com/news/jetnet-acquires-ads-b-exchange.h... https://news.ycombinator.com/item?id=34520355


Hey, I'm Kevin (founder of Airframes), and I'd like to give some background here.

While the world has been obsessing about ADS-B, which is primarily focused on positional data, I've been building a platform that is heavily focused on ACARS (often referred to as Plain Old ACARS), VDL (the successor to ACARS), HFDL (HF-based, also containing ACARS payloads), and SATCOM (such as AoA - ACARS over Insmarsat, and AoI - ACARS over Iridium).

The information in ACARS is quite interesting, and reveals significantly more about flights than ADS-B alone. The goal of the project has always been to provide a fuller picture of what is flying, including OOOI, routing, equipment status, fuel details, and more.

Some very early feeders and enthusiasts have been feeding data and supporting the effort, and as many/most HN folks know, this is all very time consuming.

Let me respond to your points too:

1. I'm a strong supporter of open source, and conceptually even open data. Your assessment that most of the stuff is still private is true. This is a strategic decision I have made. Having watched the aggregator space for many years now, my concern was that since we're still getting things built out (and I'm the only developer), a heavily budgeted competitor (such as FlightAware) could come along and enter the ACARS space long before we were ready and make the effort pointless.

I do plan to continue to open source and open data, and as it makes sense to, at my own expense, I will do so. Especially as more people become involved and build interest in ACARS.

2. The roadmap on the trello has not been updated, because, there were very few people interested in the effort and did not really engage in using it. But there is most definitely a roadmap -- some of the things are related to API, desktop/mobile app, a radio-focused OS (initially aimed at making it easy to feed to Airframes and other aggregators, but with the goal of wider radio interests) similar to OctoPi for OctoPrint.

3. Documentation has been in the works. But, seeing as that I have a lot of priorities (and it's mostly just me at the moment), it's not complete. I was hoping to release it when more digestible/ready. https://docs.airframes.io

4. I would hope that folks do consider supporting the effort, and there are a lot of big plans for it in the ACARS space.

Also, to put it bluntly, we were not yet ready for the events that unfolded. So there are some growing pains to deal with there.


Also, why not support Opensky which has an academic background and is already well established?


I want very much to believe in both projects. Both seem well intended.

Alas, for many similar reasons I'm unconvinced by Opensky as well. There are some projects up[1] but there doesn't appear to be any public development for the vast majority of the work. There are decent API docs, and ToS, which is a positive. However the rate limits are pretty low (1000 calls/day[2]) & the data does not appear to have any other distributions available.

The internet has massive amounts of invisible labor that go into keeping communities & systems afloat. It's amazing how far it gets us. But it's always unclear, how do these projects scale as needed? What happens as costs grow?

Lets look briefly at the example of another big data operator out there, npm, the javascript package manager & website/repository of packages. Npm was owned by Isaac Schlueter, who created in 2009. In 2020, IrisCouch volunteered to provide hosting services for free. They were acquired by Nodejitsu, who in 2013 ran a "Scalenpm" campaign to crowdsource some funding. In 2013 Isaac created npm, inc & began hosting the repo. Microsoft bought npm in 2020, & so far npm has continued to operate a free public registry. Figuring out how to keep this project going across a decade was hard work, with enormous sums spent & tons of work poured in. I know much less, but, like, Java's earlier pretty-big Maven ecosystem seems largely to have existed under the benevolent hand of SonaType (although other repositories have come and gone).

There's definitely some goal/dream, that projects like IPFS or Nostro or someone can help marshal & make this data available. There's still a host of other sub-problems to figuring out how something like flight tracking or package managing would actually work atop such a system. Search & discovery, authenticity, trying to effectively index/collate data to make it consumeable are all various extra levels of problem, where again centralized infrastructure plus distributed storage might work, but ideally, we could really distribute the work out. Until then, it feels like we're kind of stuck trusting benevolent for now leaders, who have limited time and hardware resources to share.

[1] https://github.com/openskynetwork

[2] https://openskynetwork.github.io/opensky-api/rest.html#limit...


The difference is adsb exchange is for adsb data, this is for all kinds of data except adsb, so the game is who adds more data sources faster


An .io service named after something plane related that's ACTUALLY about aviation? This is physically impossible, I'm absolutely gobsmacked. Are we sure it isn't some new JS framework in disguise?


I think there's something I'm missing - what's the io/aviation connection? I thought dot-io was "British Indian Ocean Territory" and was co-opted by tech startups looking for a domain that wasn't squatted in the usual .com/net/org TLDs


I think the joke is the inscarcity of tech groups that have chosen an aviation-themed .io domain to show off their venture.


Is it supposed to be aviation themed though?


Yeah I'm not getting it either. Why is this interesting vs say "Airframes.tech" or "Airframes.aero"? And there are plenty of non-programming-related .io sites too.


"Airframes is an aircraft-related aggregation service"


A lot of tech (.e.g frameworks, etc.) tend to choose aviation themes for their naming


> A lot of tech (.e.g frameworks, etc.) tend to choose aviation themes for their naming

Is it just aviation themes though? IIRC, it's pretty much any generic word with a positive connotation, with little care if that word has already been used for another unrelated framework.


The tld for aviation is .aero.


Not available to everyone though.

> Only members of the civil aviation community, incorporating airlines, airports, companies, organizations, associations, government agencies and qualified individuals, may register a domain name in .aero.


The (formerly) official discord group has cut ties with ADSB exchange and is suggesting everyone go to airframes after the recent acquisition. Don't know much about airframes or why they picked them yet though.

https://web.archive.org/web/20230125164809/https://www.adsbe...


Probably related to the acquisition of ASDB exchange by Jetnet: https://news.ycombinator.com/item?id=34520355


Sorry, I meant I don’t know much about airframes.io or why they chose them, edited my comment to clarify.


ACARS can be more interesting than just the positional data which FR24/ADSBExchange plot. Plain-text comms. between pilots and their company, ATC, etc.


That's right! ACARS has some very interesting information in it.

Just a few things of notable interest:

- realtime routing plans/changes

- OOOI (out/off/on/in) revealing when an airframe takes off or lands

- departure, destination airports and gates

- weather and other flight bag data

- airline / flight crew communication

- flight crew / airport communication

- positional data (which can be useful to augment or verify when ADS-B data is not available)

It has been a time consuming process to do research on some of the extended text (see: https://github.com/airframesio/acars-message-documentation), but we hope to keep at it.


Is it just me or is the live message feed going ballistic?


They probably can benefit from a) stop auto-scrolling when user is scrolling b) having a "throttle" slider to actually allow reading before a message zooms by.


It's not just you.


I'm glad they chose to use transparency in the drop down menu, so that I can still see the live feed going crazy behind it. Makes in to sort of a game to try and navigate. No good if you have epilepsy, I imagine.


I am not sure why they thought it was a good idea to include a live feed like that. 1000's of stations sending in 100's++ messages a second ...


It was probably fun and useful, the first week they built it. Is this an example of catastrophic success?


This is correct. It was there to show some value early on, but a complete redesign has been in the works for a couple of months now.


So when I joined ADS-B Exchange as a feeder, it was an Inc., owned by a single individual. Airframes.io is also owned by a single individual.

Make your own conclusions.


I've said it in another comment, but I believe Opensky could be the most trustworthy alternative. This is driven by academics so the likelihood for them to sell out seems a bit lower.


But you will not get data access easily.


Guessing this is posted as a reaction to adsbexchange getting bought.

Is this where people are moving?


My exact question as well…where is the next one?


https://www.adsbhub.org/

https://opensky-network.org/

https://plane.watch/

Some alternatives to adsb exchange.

For people who just want to look up planes, opensky is probably best. For people who want to feed their data to a different place, the answer is more difficult.

Please see https://news.ycombinator.com/item?id=34520355 for a more detailed discussion.


I have an RTL SDR that I haven't used yet. Can anyone recommend good Linux software for this? And is it possible to do ACARS and SATCOM at the same time on one device?


https://github.com/sdr-enthusiasts

If you want to go overboard and do it with docker - this is as good a starting point as I can recommend.


I used to have fun tracking flights over my previous home using a little usb rtl-sdr antenna attached to a raspberry pi and leveraging the dump1090 software. I moved homes, and got too lazy to re-set things up...but that dump1090 software did the trick back then. Its been a few years, so there could be better software nowadays. dump1090 was a bit of work to do the initial software setup, but thereafter was easy. The 2 versions that i recall back in the day...

https://github.com/antirez/dump1090 (original i think?)

https://github.com/MalcolmRobb/dump1090 (the form?)

Have fun!


FWIW A site that listed exactly which Airlines Owned which airplanes and the model and tail number of the airplane would be super useful to me personally





Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: