Ancient US air traffic control systems won't get a tech refresh before 2030 The Federal Aviation Administration's (FAA) air traffic control (ATC) systems are perilously out of date, but don't expect replacements anytime soon, says the US Government Accountability Office (GAO). In a report released Monday, the GAO said that 51 of the FAA's 138 ATC systems – more than a third – were unsustainable due to …
It really is an outrage
- Bluetooth <v5.4
- Insufficiently monetised to boost supplier margins for acceptable CxO bonuses
- Not written in Rust
- Not using public cloud
- No adtech API
- Probably uses a database or something in place of blockchain
- grossly excessive uptime expectations
- full of obsolete logic IC's that you will still be able to buy in 10 years time when it should have 3nm Broadcom chip du jour
I'm all in favor of of "if it ain't broke ...". However the article is reporting a report : In a report released Monday, the GAO said that 51 of the FAA's 138 ATC systems – more than a third – were unsustainable due to a lack of parts, shortfalls in funding to sustain them, or a lack of technology refresh funding to replace them. What are the parts? Vacuum tubes? Floppy disks? Windows 3.1 [https://www.ebay.com/itm/204277163690]? It would be nice to have the juicy details.
If the FAA is still using 5081's, these are specialized versions of the 3081 mainframe (370-XA architecture) from the very late 1970's / early 1980's. They were 2-way multiprocessors (2 cores in today's parlance). The technology was bipolar transistors (not CMOS) running inside of Thermal Conduction Modules (TCM) which conducted their heat to a circulating cold water system. From what I understood at the time was that IBM warehoused some number of TCMs for replacements and it is likely that they are now running out. The technology to build new TCMs is long gone. The 5081's were surrounded by specialized gear to interface them to the various ATC inputs. If these systems are still running today, it is a testament to how well they were designed, built and programmed.
> If these systems are still running today, it is a testament to how well they were designed, built and programmed.
And either somebody originally over-specced them by 10,000% or they are struggling to manage 100x the traffic volume they were built for and we are just lucky that we haven't had any oopsies
originally over-specced them by 10,000%
IBM's original value proposition was the quality and stability of their associated systems: Buy this overpriced mainframe that we've oversold to your board of directors, and you may also buy this magnetic tape, card, and printer equipment that doesn't stick, jam or breakdown, with a user interface and API that someone has tested before implementing.
I was sort of thinking that too.
More of: if the systems are working and logically would continue to work and the problem was with the lack of hardware components, that the cheapest solution could be to (re)start manufacturing these antiques. You could probably automate much of the manufacturing with modern robotics.
I supposed even the displays are CRT so vacuum tubes aren't too far from the truth.
The purely imaginary picture I have of these control centres is a scene from a steampunk graphic novel or the move Bazil. That movie perhaps prophetic.
I gave you an upvote because I agree with the sentiment that many/most IT industry ideas of "modern" are just plain silly. But really,the problems with Air Traffic Control are deeper and more fundamental. The reality: The system(s) have to combine inputs from many sources in many formats -- the aircraft themselves, ground radars of many sorts, pilots, ATC operators, weather information, many of those in multiple version (e.g. civilian, military(multiple autonomous agents), etc, etc, etc ... And some information is, of necessity verbal, not digital. And they have to do it in real time. AND THEY CAN NOT MAKE MISTAKES -- EVER.
It's a tremendously difficult problem. If you ask me, it's amazing that the "system" (if that's even the proper term for it) works as well as it does.
I've never worked with ATC although I did work on the SAGE Air Defense System in the early 1960s which had to deal with some overlapping issues. But it's been my impression that the ATC system(s) have always been in dubious condition and that it is something of a miracle that they work as well as they do.
I'm far from sure that we even know how to "fix" ATC even if there were adequate funding and a will to do so.
I remember these sort of reports preY2K. Worst was outdated systems that were bought new that could not handle Y2K from IBM IIRC. IBM advised at time that life of IT was limited to 1999 and was unable to do upgrade. Book "Safety Last" from 1970s still seems to be accurate enough.
It seems to be a common issue for many, many years in the US, no matter who is in charge, of under-investment in general infrastructure.
Many of its major bridges are overdue for replacement thanks to last of investment on replacement and repair.
US railways because they are privately owned suffer from track quality that is a joke compared to Europe.
US Airports are old and poorly designed compared especially to Asia and the major hubs in the Middle East - again is this due to the localised operational model where no-one wants to spend money rebuilding?
It always disappoints me that such an advanced and wealthy economy puts up with public infrastructure that is seriously sub-par.
Or is it just that the US voters and taxpayers don't see why they should pay for this - an argument which is pretty well moot in other economies, the idea being that in general voters and taxpayers understand that the Government invests in infrastructure [not always wisely though e.g. HS2]
US Airports are old and poorly designed compared especially to Asia and the major hubs in the Middle East - again is this due to the localised operational model where no-one wants to spend money rebuilding?
The two largest airports in the UK have 2 and 1 runways respectively, even backofbeyond regional airport in the US has at least 7.
"The two largest airports in the UK have 2 and 1 runways respectively, even backofbeyond regional airport in the US has at least 7."
Well, so long as they can afford to maintain them, that's fine. Not sure the US track record on infrastructure says they can.
But that's concrete and tarmac, sticking to the topic of ATC, not only are the FAA stating that the US has failed to budget for asset management of ATC systems. This is either a complementary or additional risk factor to the string of recent near misses due to ATC failings at US airports. They've been lucky time and again, but crap or absent systems, lack of ATC staff (and stress and overwork as a result). these problems are well known and seem to be increasing, yet for some reason the world's largest economy seems incapable of doing anything about it.
A recent example that could have been spectacularly bad:
https://www.independent.co.uk/news/world/americas/jfk-airport-air-swiss-near-collision-b2533559.html
And unfortunately a news search turns up hundreds of avoidable near misses in the US.
Sooner or later the luck will run out, and all because nobody in the entire US has the balls to limit flight numbers until the capacity backlog of controllers is fixed, and to add $10 to every single flight to fund an urgent overhaul of ATC systems.
The root cause is surely the fact that people want to fly for a few dollars and the airlines try to drive down costs across the board - they lobby, they threaten regional hubs with taking their business down the road - and then they complain about everything they have to spend money on. Over here Ryanair's boss has recently called for NATS to increase the number of air traffic controllers so they are less prone to reduced service if a lot go sick. NATS could easily* increase its controller staff, but it would put its fees up and then O'Leary - who has complained loudly and often about how much NATS already gets paid - would whinge again. He might be able to make the numbers up by selling lottery tickets but infrastructure providers cannot. Most passenger airports in the UK are shopping centres with the major inconvenience of an airport attached to them and if the airport operation breaks even then they consider themselves grateful.
*it's not actually that easy in some positions because the "spares" still have to maintain currency hours and you can't just replace one ATCO with any other - they have to be trained on that particular position.
"Sooner or later the luck will run out, "
And when it does, I suspect there will be an immediate over-reaction due to 9/11 and US airspace will be shutdown entirely until they figure out that it was just a "mistake". The the sue-balls will fly.
FWIW -- Wikipedia has a list of the world's busiest airports. I looked up the number of runways at a few of them. The only one with 7 runways is number 3 DFW (Dallas-Fort Worth). The busiest -- Hartsfield in Atlanta has 4. The second busiest-Dubai has 2.
Number of runways isn't necessarily the metric to worry about.
An airport in the middle east with no weather and 2 runways that only has A380s arriving every 5mins - can handle a lot of passengers
While a west coast airport with wind/storms/fog, local nimbys that demand you can't over fly their mansions at certain times, are surrounded by 4 other airports and 6 military bases and have to handle a mix of business turborpos, 737 and A380s arriving and departing - might have more capacity issues even with a dozen runways
This post has been deleted by its author
US Airports are old and poorly designed compared especially to Asia and the major hubs in the Middle East - again is this due to the localised operational model where no-one wants to spend money rebuilding?
Don't know where you're from (although the (mis-)spelling of "localised" kinda gives it away). But you might want to look at O'Hare International Airport (ORD) in Chicago. That sucker has been under a state of (re-)construction since the 1960's, with no foreseeable completion date specified.
US railways because they are privately owned suffer from track quality that is a joke compared to Europe.
They've heard of "single payer" medical systems like the NHS, but why would they want that? Having a railway system where individual bridges are owned, certified, maintained and charged out works for the railway doesn't it?
"The majority of the Federal Aviation Administration’s (FAA) budget comes from the Airport and Airway Trust Fund (AATF), established by the Treasury Department in 1970 to support American aviation infrastructure. It collects revenue by taxing domestic flight tickets, international arrivals and departures, air cargo, plane fuel, and travelers’ purchases in loyalty and frequent flier programs."
https://usafacts.org/articles/who-funds-the-faa-you-whenever-you-fly/
I assume that every time the FAA proposes increasing the tax levels then the airline lobbyists come out in force.
The problem is the 'F' in FAA. So one side of the house aren't going spend more (cos government spending is communism) and are going to block any spending by the other lot cos that would let them have a political win.
Ironically having ATC privatized in Europe works a lot better for society. If the airlines want better ATC because their spreadsheets say it will save them money on fuel/delays then they can just write a cheque
COBOL is easily the most readable computer language I've ever encountered. Seriously, I think most programmers could, with no training whatsoever, pick up the listing of a COBOL program and follow the logic. I did that once or twice many decades ago. On the other hand, I've never had to program in COBOL. I expect that doing so is mind-numbing.
The thing is that "COBOL" is usually used as a shorthand to mean COBOL/CICS/IMS or COBOL/CICS/DB2. CICS in an IBM teleprocessing monitor and IMS/DB2 are respectively hierarchical (yes, that ancient) and relational database producs. The "COBOL" is liberally interspersed with embedded CICS (and/or database) commands that are translated by a preprocessor that can either be standalone or integrated into the COBOL compiler. Using CICS also means that most of the traditional COBOL I/O statements are unavailable.
So COBOL as it is usually practised may contain very little actual COBOL.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
