They have 6 weeks and 1 day.......
.....to sort out that sh1t
or I'm.......actually I want to visit the states, so I had better not finish that sentence
The Federal Aviation Administration has claimed a major glitch that grounded dozens of flights last week was caused by a Cold War-vintage reconnaissance aircraft. It claimed that a U-2 "Dragon Lady" flew into airspace controlled by the Los Angeles Air Route Traffic Control Center in Palmdale, which uses a system called En …
You are quite correct...although you don't have to mention you read it elsewhere, the Reg article had that exact same information in it.
"The U-2 filed all the proper flight plan paperwork and was conducting its operation in accordance with those filings."
"Try filing an accurate flight plan next time, spooks"
Try reading the article that you are supposed to be writing sub headlines for next time, ElReg.
I don't think that it was "it forgot".... it's more of a case of conflicting rules. The U2 was a 60,000 feet and flying under Visual Flight Rules (VFR). VFR rules in the program state flights have to be lower for that. So, it was trying to get all the other flights out of the way for this bird to drop down to 10,000 feet and amount of that overloaded the system. Change one flight path in that crowded environment and there's a ripple effect that hits one hell of a lot of flights. More than the system could handle at one time.
I suppose the fix was to write something similar to "VFR is 10000 feet except U2's flying at 60000 feet".
This isn't a problem I have dealt much with, but I imagine that such programming problems as this routing is fairly hard to predict and thus a cascading complexity of routing can easily occur when it is met with extreme cases. So far so good, or rather bad, but moving on. But am I correct in my assessment* that good programming practice for such a critical system as this would not to be, reboot and try over, but rather to track the individual rerouting problems and count how many steps they would invoke. Then have a break at n steps for then to simply flag it for a human controller to deal with.
It seems rather alarming that such a critical system can cause itself to crash just because it is met with a task that it is too hard for it solve. The system thinking that it needed to be routed down to 10 000 feet is a simple goof. It not being able to solve it is a flaw. That error causing the whole system to go down sounds like a critical system failure. Worsen by the fact that it was not actually solved by the reboot. It could have easily been stuck in the mode as long as the U2 was overhead.
*I am assume here that somewhere there has been a simple filing error. Either by somebody punching in the numbers or an automated form or something. That the flight should never been marked VFR and thus the system should not try to route it accordingly. (Or something similar to this)
Easy to imagine an emergency situation in a normal airliner where a rapid descent is needed from 35,000 feet through multiple layers - Decompression, fire, multiple engine failures etc. Is ERAM going to go on strike EVERY time it's really needed?
ROTM icon - because it might be more sinister than just bad programming ;-)
A little late, but in case you read it.
Emergencies should be handled by an operator regardless, and I imagine if it had been flagged as an emergency the system would cope as it would widen the rules for rerouting other planes in the sky.
I did not say it should go on strike, I said it should hand over things it can't handle to an operator, which is more or less the whole point of a air controller system. I prefer that do rebooting in what could easily end up in a perpetual loop of reboots until the sky is clear of planes since they have fallen down.
Sidenote: Not that I read the report all that thorough, but the system is actually producing false emergency flags for operators to handle. Four way collision alerts on what turns out to be two planes and at completely different altitudes. So if they cut down on the false ones, I am sure the operator can handle a once in a blue moon occurrence where it is unable to cope.
"ERAM is a $2.1bn scheme to overhaul the systems which help air traffic controllers manage high altitude flights. It started in 2002 and was supposed to be installed in 20 "en route" facilities by 2010. However, software problems were identified [PDF] and this goal was not met."
It is not being done under a UK PFI deal by any chance???
"60K feet interpreted as 10K feet, I hope that isn't the stench of 16bit arithmetic overflow.."
The algorithms necessary to turn a mess of echoes from many objects via vertically and horizontally rotating pulsed radar beams requires that certain assumptions are made as to what is, and what is not, a likley position and speed for an aircraft to be.
The faster you pulse the radars the better bearing and azimuth resolutions you get, but the more narrow becomes the range a that you can reliably detect before pulses start to 'overlap' and you have the possibility that your return echo could be off a target in more than one location.
Software disentangles that by indicating the more likely position...possibly helped by assuming things about pre filed flight plans.
So, nothing wrong with the spooks flight planning, just ("ERAM is a $2.1bn scheme to overhaul the systems which help air traffic controllers manage high altitude flights") a crap bit of programming in the ATC software that couldn't tell the difference between 10,000ft and 60,000ft.
Don't you guys actually READ the content of the article before writing the subheading?
The U2 was never really a 'stealth' plane. When it was designed, it's main benefits were it's high operational altitude (higher than the Russians Surface-to-Air missiles or fighters), which lulled the Americans into a false idea of it's safety, and the high endurance that allowed it to overfly most of the Soviet Union. In the years before surveillance satellites, this was the main method of identifying what the Russians were doing.
That's why Gary Powers being shot down was such a shock!
The SR71 added some stealth features, along with very high speed, which enabled the Americans to continue surveillance operations.
The SR71 also added the feature of never overflying enemy territory - at least not of any enemy with an airforce.
Over-fligths were banned after the U2 and because of a danger that somebody with an itchy button finger might mistake the SR71 for something a little more explosive
The SR-71 did overfly enemy territories with an air force and an air defense - otherwise it would have been just a nice exercise of high-speed flights. It was its speed and altitude that made it impossible to be intercepted - the reaction time was not enough for a plane - even the MiG 31 - to take off and climb enough to get into a fire position, nor for a AA missile to reach it - and it was too fast to be a cruise missile, and to slow to be a ballistic one.
The MG-25 could in theory intercept the SR, as it would be rare for the SR to being flown at full speed for long, and certainly armed with a AA-6 Acrid missile, it could take out a 71. Only problem is the Foxbat may be completely knackered afterwards.
All theory of course....
The SR-71 flew many times over enemy territory. Quite a distinguished career. According to wiki, a fifth of its flight time was spent at Mach 3 which goes to say something. There arent many vehicles spending a fifth of their time at the red line so that says something about the engineering. Also not a single one was shot down so the concept worked, noone could catch it simply because you would never be able to accelerate fast enough to get to it, unless you knew the flight plan beforehand and were loitering.
If it had a data link then it would have been even more useful.
The SR-71 was designed to fly at its highest speeds for most of its flight. When cold, on ground, it even leaked fuel from the tanks because the airframe was designed to "seal" only when the high temperatures reached during flight expanded the metals. Speed and altitude were its only defensive weapons, and that means the plane should have been already at its planned speed well before entering the enemy territory, and start to slow down only after it left it well behind, to ensure the air defense had no enough time to react to the approaching or leaving plane.
Sure, a fast plane like the MiG-25 could have tried to intercept the SR-71, but not if he had to scramble from ground and reach an altitude where it could fire its missiles. The MiG-31 is even more capable, but yet it needs to be already in flight and in a good position to have a chance to intercept an SR-71.
Powers being shot down wasn't a shock. They knew it would happen sooner or later.
If a missile didn't hit the thing, simply evading it could rip the wings off (they fly in what's known as "coffin corner" where cruise speed, stall speed and VNE (Velocity to Never Exeed) are all within a couple of knots of each other. (It's basically a starfighter with extremely long wings)
The real shock was that the integrated destruct system didn't work (and that the pilot survived/didn't use his suicide pill). The aircraft was rigged with high explosives along its entire length for just such an event, to ensure the Eviiiil Russkies didn't get their hands on a complete example, but the G forces Powers was subjected to left him unable to hit the switch before he bailed out.
Back in the 50s the USSR has better radars than the USA. The U2 was designed under the assumption that it would escape detection by flying at 60 000 feet. They realized the error when they saw all the pictures taken were full of MIGs following the U2.
They tried to increase the altitude, but quickly gave up.
Interesting. I did my flying training in the 1990s at Long Beach, just a short distance from LAX, and a fellow student claimed to have heard the following exchange over the radio (with civilian ATC).
Pilot: Air Force XYZ requesting flight level 600.
ATC: Man, if you can get up there, it's yours.
Pilot: Air Force XYZ currently *descending* though flight level 870.
I never knew whether to to believe the guy or not. FL 600 (i.e. 60,000 feet) and above would be Class E airspace, not requiring explicit permission to fly in it.
That episode is in (I think) Francis Spuffords book 'Back Room Boys'.
the SR71 pilots tell of how they are indeed at 80000 feet over Cuba, wearing pressure suits.
They look down on a Concorde below and reflect that the Brits/French have engineerd an aircraft which can fly up there with them, at high Mach, with people inside in shirtsleeves drinking Champagne.
Oh, and if you haven't read it, Ben Richs book 'Skinkworks' is fantastic
The U2 is indeed a Starfighter body with long, thin wings.
That sounds like one of the 'Aurora' conversations. Several conversations between unidentified aircraft and ATC along those lines i.e. pilots alerting ATC that they were _descending_ to a Flight Level way above the capability of known aircraft have been reported by various plane spotters and cited as evidence for the Aurora spy plane. Other variations include F-4/F-15 jocks reporting ascending to FL600+ to ATC in a boastful sort of way only to be 'trumped' by an Aurora pilot reporting that he's descending to an even higher FL.
None of these conversations can be verified though, so they don't really count for anything.
On the other hand, there has been some pretty good evidence for an Aurora type aircraft. Amongst the best evidence is a sighting over the North Sea by someone who had been in the Royal Observer Corps International Aircraft Recognition Team. Also, a series of sonic booms recorded by the USGS seismic sensor array in Southern California which, when analysed, indicated an aircraft, smaller than the Shuttle, flying overhead at ~90,000ft @ Mach 4-5. Then there was a photo taken by a geosynchronous weather satellite that appeared to show a very high-altitude, high-speed contrail starting at Groom Lake and extending directly East across the Atlantic Ocean (it had to be created very quickly, i.e. by a very high-speed aircraft, otherwise it would have started to disperse at the start of the contrail). Not sure if this photo was ever verified though. Most of this evidence dates from the late '80s, through the '90s to the early '00s. However, there have been some reports of more recent sightings over Kansas and Texas, with photographs, earlier this year, in February and March, which might tie in with Aurora missions associated with the on-going Ukraine/Crimea situation.
Probably the best evidence against an Aurora type aircraft is that if it does exist then it won't be entirely unknown to the military forces of the rest of the world, even if they don't know its full capabilities, and the only people to whom it's existance is actually being kept secret are the general public, who don't really count, so what's the point in keeping its existance secret from the public when the rest of the world's military know about it?
>However, there have been some reports of more recent sightings over Kansas and Texas
This is believed to be a B-2 successor/big brother. There also appears to have been a laser test of significant magnitude. The tie-in to Crimea/Ukraine would be that this 'unveiling' and 'light show' was a warning that though the US doesn't want a war, it could certainly win it.
>so what's the point in keeping its existance secret from the public when the rest of the world's military know about it?
Keeps the other side guessing as to the true capabilities. The deterrent value may be greater than its battlefield worth. Revealing it quietly is sure to get China and Russia's rapt attention.
The pertinent bit
"One day, high above Arizona , we were monitoring the radio traffic of all the mortal airplanes below us. First, a Cessna pilot asked the air traffic controllers to check his ground speed. 'Ninety knots,' ATC replied. A twin Bonanza soon made the same request. 'One-twenty on the ground,' was the reply. To our surprise, a navy F-18 came over the radio with a ground speed check. I knew exactly what he was doing. Of course, he had a ground speed indicator in his cockpit, but he wanted to let all the bug-smashers in the valley know what real speed was 'Dusty 52, we show you at 620 on the ground,' ATC responded. The situation was too ripe. I heard the click of Walter's mike button in the rear seat. In his most innocent voice, Walter startled the controller by asking for a ground speed check from 81,000 feet, clearly above controlled airspace. In a cool, professional voice, the controller replied, ' Aspen 20, I show you at 1,982 knots on the ground.' We did not hear another transmis sion on that frequency all the way to the coast."
The (fascinating) book "SR-71 Blackbird: Stories, Tales, and Legends" by Rich Graham documents an exchange similar to the one you describe. The air traffic controller was reported to have replied by saying something like: "How exactly do you intend to get to 60,000 feet?", to which the pilot responded: "Actually, I wish to descend".
They aren't flying Vulcans because they are made from a magnesium alloy. It rots and limits the life of the airframe. They'd become too expensive and require too much maintenance to keep flying, plus the engines are no longer in manufacture so spare parts would become an issue also.
Perhaps it's not a glitch at all! The discrimination algorithms would normally shunt the U-2 flight into the UFO traffic control system. However, because the U-2 had correctly filed a flight plan with a military designation the UFO traffic control system kept kicking it back to the regular traffic control system that looped it back to the UFO system.
Talking to a controller at NATS a few years back when they were introducing their electronic system, I gathered it also made the handling of a lot of the US-bound transit traffic from Eurocontrol through to the Atlantic a lot simpler as that traffic was already at cruising altitude and didn't usually need a lot doing to it, so the system could accept the "digital strips", effectively doing the paperwork and just let the controller keep an eye on it, focussing on sorting out those flights that actually needed at ascend/descend from/to airports or lower flight levels.
As he said though, it's nice, but when you have a problem, like an erroneous number on a strip, you can't physically throw it at an ATC assistant and have them figure it out, you've got to work out what's wrong, and manually correct it yourself, which distracts you from the rest of your airspace.
That said, errors shouldn't occur (as often) because the data transfer is automated, so no mistaking what the lovely Belgian controller said through their accent, and there were obviously failsafes that flagged for attention if you were asking an aircraft to achieve orbit or go Mach 7.
It perplexes me that they wouldn't have written in exceptions into the code for cases like this. Such an application should be able to detect the amount of resources a calculation is taking and kill it if it gets to a certain point and informing a human that that particular aircraft needs to handled manually, with the option of completely ignoring it. Failing like it did is completely unacceptable, even for the most basic of commercial application let alone something with safety risks.
But I wonder, what other bugs are still in the system? What will happen if a transponder malfunctioned and started pushing out false data? Would it crash again if SpaceShipOne/Two were to fly overhead? What kind of testing was actually performed?
COG: "...an application should be able to detect the amount of resources a calculation is taking and kill it if it gets to a certain point and informing a human that that particular aircraft needs to handled manually, with the option of completely ignoring it."
You mean like a "1202" Alarm?
The 1202 alarm was caused by the fact that Neil (Or was it Buzz?) had forgotten to turn off some unneeded ranging systems on approach for landing, which those active systems where interrupting the CPU too many times for it to operate in real-time and needed to drop instructions to get back to the business of planting humans on the moon.
Yes, exactly. Given the nature of this application, especially since it is critical that its real-time, you should have monitoring and reporting systems to warn the user that bad things might happen so they can be prepared. In the case of the 1202, it was simple matter of toggling a few switches to get everything back to ship shape for the return flight and future landings, in this application they could have debugged it or added more resources.
Reading this, I was reminded of the first time I gave my sister a 'driving lesson' in the front driveway.
So she wouldn't stall, I had her over-revving the engine before slowly releasing the clutch.
However, as soon as the car started to move, she lifted both feet up, held her hands in the air, and screamed "I want to stop". However, due to the high revs, the car didn't stall, but jumped forward until hitting a nearby wall.
It seems that this system had a similar panic attack. But you'd expect it to react better than a 16 year old kid who's only real worry was what my dad would say when he saw the car (and the wall!)
"However, as soon as the car started to move, she lifted both feet up, held her hands in the air, and screamed "I want to stop". However, due to the high revs, the car didn't stall, but jumped forward until hitting a nearby wall."
Sorry, I hope you were both OK after that, but that made me LOL.
"Sorry, I hope you were both OK after that, but that made me LOL."
:-) Yeah, a slightly dented wing and a few stones knocked from the wall were the only casualties (apart from her pride!)
It's funny - this happened about 25 years ago (showing my age!) and I haven't thought about it in years, but this story triggered the memory of someone/something throwing everything in the air and shouting "I want to stop!"
I'm not buying it. Unscheduled and improper flight activity happens all the time, and is dealt with procedurally and with minimum fuss.
the area has handled the outrageous flight performance envelope of the SR71 crossing over it's airspace without issue for decades. If something with unplanned flight path and unpredictable performance behavior were to crash the system, it would have been that. Most flight controllers can tell you stories of aircraft whose responses to orders to "make room" for emergencies, can suddenly stop pretending to be an airliner transponder and behave in ways most aircraft enthusiasts can recognize from their copy of Jane's.
This is something else.
Is it a coincidence this comes in at approximately the same time as reports of Russian aircraft patrolling near the California coast and requiring Air Force intercept?
Biting the hand that feeds IT © 1998–2019