Floppy disks and paper strips lurk behind US air traffic control The Federal Aviation Administration (FAA) has confirmed that the US air traffic control system still runs on somewhat antiquated bits of technology, including floppy disks and paper strips. This came during last week's Budget Hearing for the US House Appropriation Committee, in which the current FAA boss, Chris Rocheleau, …
The latest generation of Cloud or on-prem systems is useless if they don't stop firing the experts who run those system and, more importantly, can act when the system goes down.
The next time one of Elon's boosters fails to make it to orbit, you might need those same people to clear the airspace underneath it.
What happens if the aircraft crash into the computers that are supposed to be controlling them? Surely cloud-based ATC has the advantage of easier control, but the disadvantage of having to dodge out of the way every so often...
I think ATC should just shout at the pilots more. Tell them to slow down, be more careful, look where they're going. Stop showing off! You'll fall and hurt yourself. That sort of thing. More Mumsnet, than Skynet.
Assume that was said in jest. ATC systems don't run on ordinary Windows PCs and servers.
It's all specialised kit that communicates with physical hardware at the airport, on dedicated links.
For example, the airports have dishes that pick up transponder signals from inbound aircraft so ATC can "see" where a plane actually is.
Trying to put that in the cloud would cause deaths. That is not an exaggeration.
Paper strips cannot be hacked from the other side of the planet. They don't require power to be read. Digital is inherently less resilient than most manual, paper and analogue alternatives. Something to consider as the climate degrades and politicians restart the cold war to take back control of their citizens.
AFAIK there are indeed strip printers, but their use has remained the same for literally decades because it just works. The strips get inserted in holders which get put on a sloped board and so show the queue. If a plane gets moved to another controller it's a matter of moving the holder to their queue. Once off queue the strip gets removed and the holder recycled. I don't know what they do with the strips - I never noticed (as a kid I was very often watching in the actual tower of our local airport - yes, that is long ago :) ).
Bonus: strips work without power once printed. And even before if written, but there's a dependency on the quality of the handwriting used so no former GPs :).
"no former GPs"
Given a good proportion of defrocked medicos were given their marching orders for their attachment to prescription psychoactives, I am not sure you would want the spaced-out ex-GPs controlling your airspace.
Those given the boot for interfering or otherwise forming unprofessional liaisons with their patients obviously had difficulties in keeping their mind on the job in one domain; I don't imagine they would be more focussed in a ATC tower, ;)
"The strips get inserted in holders which get put on a sloped board and so show the queue."
Wow that's very old air traffic control. I watched a show about two planes colliding over the US Grand Canyon in 1956.
Queues on a board is what they were using then. IIRC, Updated every hour when each plane radioed in with altitude, speed and which route they were on. They had radio beacons on the ground so they could stay on route. They radioed their airlines which relayed the information the the controllers which kept a board for each route and altitude and last known position. And they cleared changes and relayed instruction back the airlines.
Not as many planes in the air back then but it was just barely working by 1956.
IIRC, the crash happened because the planes were all flying visual rules and had a bit of leeway in what they were doing. So the two planes ended up at the same altitude because of clouds, one plane was given permission to fly above the clouds, so ended up flying the same altitude, but hadn't needed to report back yet. And then due the Grand Canyon most planes would loop around to let everyone have a good look. So the planes were on the same route, but the one had left a few minutes earlier. But with the sightseeing, the one ended up coming up behind the other plane. I forget why the following plane didn't see the first.
No black boxes or survivors. But very good in investigation of the wreckage lead to figuring it out.
The investigation set the standards for later plane crash investigations. And the crash also led to updating the the air traffic control system.
“I forget why the following plane didn't see the first.”
Both pilots were looking down at the Canyon and oohing and aahing. Everyone knows that’s the job of the pilot flying, while the other pilot keeps vocalizing the estimated distance to the other aircraft. Or just screaming if they get too close.
Sort of. It depends. I've lost a lot of paper stuff while moving, and a lot more is extremely hard to access because I wouldn't know whether it's in my parents' attic or in my attic, and wouldn't know in which box it is, and it's all unsorted. On the other hand, I still have a whole lot of digital documents from when I was a kid, they are actually accessible, and it's all backed up to an encrypted could storage. You could probably hack or disconnect me if you tried hard enough, but if it was something running an airport (instead of old D&D campaigns) it would be air-gapped and offline and have UPS and a gennie.
Arguably, the objection to that would be that I was clearly not storing my paper properly while at the same time being decent at storing my bits and bytes, and it would be absolutely true, but that's the whole point - the thing that gets you isn't digital vs. paper, it's doing the job well vs. doing it poorly.
"wouldn't know in which box it is, and it's all unsorted."
Reminded me of long ago when digitising hand drawn maps and engineering diagrams, manuscript notes were still in the future. This material was was stored in cardboard record boxes labelled with a geographical location reference.
When PCs became available (actually CP/M-86 in this case) and Ashton-Tate dBASE III an inventory of each record box was entered into dBASE which then could be queried to locate the box(es) relating to particular owner or client, property address, type of work etc.
A glorified library card index system but somewhat more versatile but the key feature was that the original record system remained intact.
"On the other hand, I still have a whole lot of digital documents from when I was a kid, they are actually accessible, and it's all backed up to an encrypted could storage. You could probably hack or disconnect me if you tried hard enough, but if it was something running an airport (instead of old D&D campaigns) it would be air-gapped and offline and have UPS and a gennie."
The paper strips used to manage flights is very ephemeral. It's life is measured in minutes. Once the aircraft moves beyond that ATC station's domain, a new strip appears in the next domain's queue.
It's questionable how much flight data needs to be archived once an aircraft completes a journey. There's a couple of YouTube channels I watch from time to time presented by commercial pilots that fly small craft. On a long journey, they might check in with a controller, fly some more, and get a check out call as they are passed along to the next patch of airspace. Other times, they are getting all sorts of instructions on the way up/down, getting permission to swerve around a bad patch of weather, etc. Even at those busier times, how much of that needs to be archived on a regular basis?
The work flow from the old system is a good start and should be something that isn't difficult to replicate digitally since it's well defined. To insist on a whole pile of "improvements" in one go is begging for problems. It's pointless to try and improve something that already works. ATC has become what it is today by fixing serious inadequacies and not by reworking a UI for " a better user experience". The visuals on aircraft stuff is extremely old school, but I got used to it quickly enough. I can look at a section map and interpret what's going on with little processing. I do like the digital versions which are substantially the same, but I can strip back layers of stuff I don't need for my flight based on my flight plan. If I'm way up in the flight levels (over 18,000ft), the map gets really basic. The idea was to not completely replace the map, but to give it options based on needs. I still have paper maps for all of the places I fly and take them with me. The lack of batteries to go flat is a big deal.
The thing with paper is that he information is only available to the person holding it. Twenty years ago at Copenhagen airport they would point camera's at the strip bays so that others who needed to know could see the clearance status of flights (the position in the bay indicates the clearance). If for example the departure sequence was changed, the order of the strips in the bay was rearranged. Any changes to the strip were made by crossing out the information and writing the new information next to it.
All of these changes were manual. With modern Electronic Flight Strips (EFS) all the information of who changed what and when is logged and available for analysis. Everybody who needs to know can see the electronic strips.
Unlike paper strips, electronic flight strips are presented in color, making information much easier to read. Arrivals and departures are much easier to tell apart.
There often is also a direct link to the traffic displays of the Advanced Surface Movement Guidance and Control System (A-SMGCS) so that the information labels of the aircraft are always in sync with the information shown in the strip. Information is kept in sync in both directions. For example, alerts are also made visible on strips in real-time for better situational awareness of the controllers.
Paper indeed does not need electricity, but any modern airport handling a take-off every 60 seconds cannot afford the inefficiencies of handling paper. By the way, usually after the flight has been handled, the strips are taken out of their holder and discarded in a special bin. At some airports the contents of this bin is taken to an administrative department where the information is manually entered in a database. Not efficient and the database is guaranteed to have errors. Overall, electronic flight strips are a big win.
Giving an electronic system a snappy name like "A-SMGCS" doesn't mean the system is reliable or resilient. The fact is that time and again new technology is being introduced to replace old, outdated, inefficient systems with systems that are inherently unreliable and unstable. Its not the fault of the system but rather that of the suppliers who just don't seem to be able to make things that stay up for extended periods. Now, if Jason Hwang (nVidia) is to be believed, suppliers don't even need to know how anything works any more, you just "program it in human" (i.e. feed it a Powerpoint and hope for the best).
Much of my working life has been with control systems, things that have to function exactly to a specification and continue functioning no matter what even (in some cases) after some smoke has been let out. So I'm not some latter day Luddite, its just I know enough about systems design to never take the glib words of a salesman (or a programmer for matter) at face value. There's just been too many disasters, so while replacing a simple manual system with a computer should be a work of a moment there's a really high probability of a total SNAFU.
Ground movement control systems are not commissioned overnight. National regulators require a safety case in which all possible aspects of a system's operation are considered, including operating procedures, environment, traffic patterns, human interaction and failure modes. The system only goes into operational use if the remaining risk is acceptable to the regulator. Back in 1998 the Internationbal Civil Aviation Authority (ICAO) deemed A-SMGCS just as good as looking at traffic out the window, that is the level of sophistications these systems attained already back then.
A typical required reliability figure would be 99.995% uptime. To properly monitor all ground traffic, aircraft and vehicles, there is often multiple coverage by radar, multilateration and ADS-B (Automatic Dependent Surveillance - Broadcast).
The systems are designed for high availability with redundant hardware spread over at least two locations and duplicated data links with physically separate pathways. The design allows for graceful degradation so that if central processing fails, the radar image is still available.
Runway access is controlled by stop bars, often with microwave barriers that will detect runway incursions. If a stop bar is off, that does not mean that a crossing is safe, but if it is on, the pilot knows it is not safe.
Moreover, Air Traffic Controllers are trained to deal with system failure. In that case, they switch to procedural air traffic control, doing everything in their heads with of course a much reduced traffic flow.
This is clearly one of those cases where the effort involved in designing, implementing, and certifying a system dwarfs the cost of manufacturing the components or operating the equipment by a significant factor. You get one shot to get it right and can't really turn it off. But as long as the system is kept to spec and properly maintained, it'll always work.
It seems like there'd be a very lucrative opportunity to start a company that does nothing but manufacture "frozen-in-time" hardware and train people how to service it. A billion dollars (peanuts in any FAA modernization scheme) could pay a thousand people $100K/year for ten years to keep it going. Again, is the US government even going to notice a billion bucks a decade? Definitely not.
Once this outfit is up and running I would very much like to work there!
This post has been deleted by its author
That raises the question of what it runs on. Anything in use for 25 years would have been designed at a time when floppy disks were seen as a reasonable way to transfer data.
If the issue is maintainability either long term availability of spares needs to be established upfront or it needs to be capable of being re-hosted on new hardware (?and OS) without losing qualification.
This post has been deleted by its author
If there were enough industries/applications they would benefit from such "frozen in time technology" you might build an entire manufacturing chain from the silicon wafers to your LSI-11, microvax, Z80, m68000, support ICs, static ram etc. The low volume and undemanding (by contemporary standards) fabrication technology would today be well within the capabilities of many University Engineering departments.
Some components like floppy disc drives and media could be replaced transparently with solid state emulators with very little effort required for recertification (given actual floppies were barely more reliable than wet paper.)
Something like this does seems to happen in industrial control & instrumentation where last century PC technology like Intel PIII&486DX motherboards with ISA bus & slots were still being fabricated and available at a price (until at least a few years ago.)
Yep, the Voyager software and hardware frozen for 50 years and still running in sub-zero conditions. Advances meantime have been great but no match for the ingenuity of today's controllers who continue to make it just work.
I just spent the weekend resurrecting some old DOS stuff. Its remarkable how good, but simple, that stuff was. Just like when I put one of my Vinyl LPs back on a turntable to discover the sound is possibly still better than the CDs responsible for archiving them out of sight and sound for four decades.
Though you have to at least entertain the possibility that in the years since you laid those LPs to rest, your hearing response at frequencies above ~ 8 kHz has dropped to nil, so you can no longer hear the white noise from needle wear and tear, to say nothing of the snap, crackle, and pop of your all-valve (tube) amplifier. Used LPs and retro electronics: perfect for the free bus fare set.
"Some components like floppy disc drives and media could be replaced transparently with solid state emulators with very little effort required for recertification (given actual floppies were barely more reliable than wet paper.)"
That was why Gotek floppy emulators were invented. Popular in the retro computing community, they were designed to replace the floppy drives in factories full of sewing and knitting machines because those machines were still doing a perfectly adequate job but the floppy disks and their drives were becoming harder to repair or replace. Likewise, USB to RS232 and other similar devices. You don't want to replace multi-million quid equipment just because the new PC that controls it no longer has RS-2323C ports on it :-)
This post has been deleted by its author
Complex mission-critical systems are very resistant to change. You will always see old tech in these environments. So it would make sense that you make provisions for the ongoing procurement of parts and maintain expertise in those systems.
While the system is obviously more fragile than it should be, I find it irritating that old but still useful tech is thrown under the bus. Perhaps it's a way to make the age of the system relatable to non-technical sorts.
- Floppy disks have worked for decades. Now, if they lack sufficient inventory of the disks and the drives, that's a different issue.
- Paper strips just work, and bring a physical component to the controller. Maybe someone will come up with a brilliant way to make this work digitally, but human factors are very important here.
- So the argument is that replacing copper with fiber prevents the line from being cut? Fiber can have benefits over copper, but if being cut-proof is one of them, it's been hiding that characteristic well.
"It seems like there'd be a very lucrative opportunity to start a company that does nothing but manufacture "frozen-in-time" hardware and train people how to service it. "
That's a really good idea. If storage isn't a big deal, building and verifying systems that will sit on a shelf can mean good money down the road when that spec hardware and configuration isn't available new any more. When an order comes in, a system is pulled and certified and sent to the customer ready to install. It time and a requirement for the exact kit is key, it can be worth good money.
A number of years ago, I recreated a bunch of PCBs that were used in high rise building fire/alarm systems. The original manufacturer had stopped making and supplying replacements years before but the buildings still stood. To redo the whole system could be millions, so a few thousand for modern replacements was a deal. I'd love to find some more of that sort of thing. It's something I can do rather cheaply these days and get most of it done out-of-house. Sure, many others could do it too, but having something on the shelf and ready to ship is important. The ones I did make were definite improvements. I could see with the returned cards where there was heat issues and what capacitors were abused the most. I never had any returns so I think I nailed it and it's the sort of thing where being perfect is better than compensating with good customer service. A cheap LED streetlight lamp is no deal when the cost to replace a failed one is many times more expensive than just buying the premium one and not having to worry about it.
I gather it now costs comparatively little to get obsolete sillicon chips remanufactured. There's company(ies) that have specialised in preserving the lithography masks, and what was once a costly shit-hot 100nm fab is now a not very busy 100nm fab that's open to any work going.
I hear that this has had an impact in some military systems upgrades. Apparenty it can now be cheaper to get old electronics re-manufactured and brought back out of the "obsolete" bin to run existing binaries than it is to stand up a software dev team simply to port the existing source code to a newer version of Linux (a job that should be no more than ./configure, make, make install).
Ok, this is not exactly "routine", and it's full of ifs and buts, but the fact that it is happening at all is remarkable. This is especially remarkable as the whole reason we have portable source code at all is partly thanks to the DoD mandating POSIX (a standard that kinda coallesced what we think of now as Unix, and prevented more OS fragmentation than already existed at the time). Except now the software compiling is the expensive bit.
This ought to be a wake up call for the software industry. It's long been the case that the myriad minor variations in software development ecosystems has inflated the costs of keeping software alive. Variety is good, but it's also bad. Every time someone decides to tweak how cmake works, it costs someone somewhere a fortune in updating their build environment (or creating a fresh build environment). These day, for anything even slightly chunky you now need a devops team simply to manage the build environment, on top of the engineering team actually writing the software. This is going to kill off software development. As tools / applications mature and become "just fine for the job", the shop who can keep that software running as built 20 years ago is increasingly going to be the winner. Cloud won't help because all Cloud does is mean "at least you're not having to maintain hardware".
Always the right time for this quote:
"There is nothing a mere scientist can say that will stand against the flood of a hundred million dollars. But there is one quality that cannot be purchased in this way — and that is reliability. The price of reliability is the pursuit of the utmost simplicity. It is a price which the very rich find most hard to pay."
It is scary and funny at the same time: YouTube link
My work often takes me to interesting places, some of which I am not allowed to speak of. At one such establishment containing several very large fuel tanks, a new fibre optic cable was laid to transmit the fuel state of each tank plus flows in and out of the system to the central fuels office, about half a mile away.
Then they decided to replace part of the site lighting system which involved digging a trench for the underground lighting cables. Unfortunately (you've guessed it) the trenching contractor severed the fibre optic link (in several places).
So for the past 5+ years someone has to physically go out to the tanks every time there is a fuel transfer to read the local gauges.
In addition, given the state and reliability of modern software, I await the first time that a software update brings the whole system crashing down taking 48 hours to reboot and restore. Let us hope that they retain the paper strip system as a backup.
Why are a lot of avionic ATC and rail CBTC systems still running on old hardware (80386s, 80186s, and in many cases 8085s or Z80s) and software (usually proprietary, but sometimes DOS based, going back to as old as DOS 3.1)?
Because they work, they are reliable, and they don't suffer the complexities of modern systems.
One customer I worked with in 2008 recently decided they wanted to upgrade their existing system. What was wrong with their current system? Nothing. It met all safety and reliability standards, the operators were trained on it, and it's been running for 17 years without any real issues. Was there any functionality they wanted added? Was something missing, or not working?
No, the supposed problem was that interface wasn't "current", and there were newer, shinier technologies available now. I mean, why would anyone want to still use Ethernet when Wifi 6 is available?
Uhh... because it works?
Outdated and obsolete tech should absolutely be replaced before it fails. But too many (as in, almost all) of the infrastructure upgrades I see are the result of VIPs going to trade shows, seeing new technology, and having the tech presenters say "you're not still using that old technology are you?", when in fact there's nothing wrong with it, and there's no issue to be solved.
There is a very real problem being dependent on hardware parts that are no longer made. But 5.25" inch floppies are still being made and sold (hell, here's the Amazon link for some), as are a lot of older tech. They may not be used by home users any more, but there's enough of a market for a lot of this older stuff to keep at least a couple of suppliers going.
Unfortunately, a lot of the drive to upgrade is simply a desire to appear current. I had a customer want replace a perfectly functional 2012 era Windows XP system with a more modern Windows 11 architecture. The XP system was air gapped, and not connected to the internet, so there were no security issues. But for Windows 11, they decided that they needed cloud access for some unknown reasons. The result was that the budget for cyber security alone for their new system was about 60% of what the original 2012 system cost in the first place.
They're spending almost as much money on their new system to prevent problems that wouldn't even exist if they weren't upgrading in the first place.
I think the first effort started during Bush I and through the Clinton years and was killed early in Bush II's term after 9/11 because now they had to add in different priorities e.g. more consideration for security. It has had a couple more incarnations and now here we are with Trump starting it all over again for what the fourth time now? Maybe fifth?
Despite his ridiculous promises of it happening in three years we all know it'll take most of that time just to figure out the requirements and award the contracts to his cronies who have no more experience in this field than his cabinet does in the posts they occupy. So it is safe to say we'll be starting over again in the 2030s.
In the meantime maybe we could paper over a few of the weak points with the current system. For example, do we really need to continue depending on floppy, considering that multiple solutions ranging from USB to floppy interface adapters to floppy image emulation software exist. Surely one of them will fit the needs and increase the reliability and/or efficiency of the current scheme. Yes yes we need time to fully test it but pick a less busy airspace and pilot it (no pun intended)
I have to imagine something that improves efficiency over the "strips of paper" could be come up with too. I won't make a suggestion as I don't think anyone could without spending the day in a busy air traffic control tower and seeing how they are used before you could even begin to come up with ideas to improve upon them. But one risk that comes to mind is handwriting - not everyone's is clear which becomes important when someone goes on break or has to leave for medical reasons or whatever. While you can "get used to" someone's crappy handwriting you don't want someone squinting at a scrawled strip of paper trying to decide if that's a 5 or an S for an extra half second when there isn't a half second to spare.
... but keep it as backup.
I think even planes with advanced avionics still have basic analog instruments as backup. They did until very recently certainly.
Altitude, Speed, Heading, actual Compass, Fuel, etc - tucked in a corner of the flight deck.
The basic stuff to keep a plane in the air and land it - in case the advanced avionics give out.
Similar approach with ATC might be a good idea.
I got systems still running from the late 70's-80's thru to the 90's generation of hardware.
This stuff is pretty much bulletproof except one thing, crt displays.
But, field monkeys like me working with video game (commercial) repair companies have figured out how to mod them with flat screens using CGA to VGA converters.
I can still source pretty much all the TTL components on board assemblies. Power supplies can be rebuilt as long as the transformers are good and Molex and AMP still make all the connectors.
Even the Z80 processors are available still.
And they keep chugging along,.........!
Not bad for 45 year old iron.
It's going to be the older the system the easier it is to keep going, just as with aeroplanes and racing cars.
Keeping an old spitfire or Lancaster flying - no problem, something just 10 years later such as the Vulcan - very difficult, 10 years later again such as Concorde - impossible.
Similarly with racing cars, any competent mechanic can keep an F1 car from the the early 20th century to the late 80s going, but after they started replying on bespoke ECUs you need a team of specialists from the originating manufacturer to even get them to start. By 2008 the FIA mandated a standard single ECU used by all F1 teams, but there are custom chassis and engine apps, and even if you know how the TAG-320 works, without access to the thousands of parameters which are loaded in, you aren't going to get very far.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
