Forum adverts like this one are shown to any user who is not logged in. Join us by filling out a tiny 3 field form and you will get your own, free, dakka user account which gives a good range of benefits to you: No adverts like this in the forums anymore. Times and dates in your local timezone. Full tracking of what you have read so you can skip to your first unread post, easily see what has changed since you last logged in, and easily see what is new at a glance. Email notifications for threads you want to watch closely. Being a part of the oldest wargaming community on the net. If you are already a member then feel free to login now.

A question for the hive mind, has anyone seen any serious analysis of the question of whether manned warplanes have a future?

Looking at the ever-rising cost of the F-35 and how ubiquitous drones have become in reconnaissance and (whisper it) assassination I wonder if the next generation of fighters and bombers will even have crews.

The never ending debate about the A-10 revolves around how dangerous its mission is (flying low and slow and shooting ground targets) so a drone would be a natural replacement. And if attack planes are replaced can drone fighters and bombers be far behind?

Aircraft carriers replaced with drone launching subs?

In fact I can see the bomber being replaced entirely by cruise missles with submunitions which already do a lot of the hard work of pounding air defenses.

Basically if you were planning the navy/air force of 2068 would you see any role for manned warplanes?

The aviation industry has been debating this for about 20-years, I think the best answer at the moment is "maybe?"

It's actually a lot easier for pure war-fighting scenarios; pounding enemy air defences, destroying armoured columns, etc. is fairly straight forward to programme in.

However full-scale war is actually (fortunately) not very common these days; your aircraft are much more likely to be policing no-fly zones or intercepting airliners that have lost communication or drifted of course.

The situational awareness required to fly alongside a 787 and take a judgement as to whether it's been taken over by terrorists or not is very hard to achieve with a drone; you're much less open to error if you have a highly-trained professional on the spot.

It's the same reason that beyond visual range missiles have never lived up to expectations, even now the technology is accurate and reliable. The likelihood of having Rules of Engagement that allow you to splash targets at 20+ mile range is very low; visual confirmation is the norm.

All of that is before you get into concerns about cyber warfare; jamming or hacking drone control systems. Become too reliant on drones and your entire military could become neutralised (or worse case, used against you). And totally autonomous vehicles, not reliant on communications, is a step I don't think anyone wants to take (Skynet anyone?!)

I think the most likely scenario is manned warplanes acting as local command and control nodes for a local "swarm" of semi-autonomous UAVs; able to fly themselves, but receiving targeting direction from a human.

The point about needing people in aircraft to communicate and police no fly zones and the liken is a good one. Also everyone is investing in electronic countermeasures to drones to hack them or cut their communication lines. I read a theory somewhere that if the ability to disrupt electronic warfare becomes advanced enough, troops will have to fight the old fashioned way because nothing computer based will be reliable.

I don't think we'll ever see fully robotic controlled armed forces in our lifetimes.

Even if we can build AI which are fully capable of controlling every aspect in real world situations, there is still a huge trust element. People will want a human mind to be at least overseeing if not there in the field to react to the real world situation. Even if a machine can be programmed to react the very same and a lot faster.
Furthermore the human pilot would be a back-up for the machine.

I could well see a controller type aircraft with support drones being a viable approach toward this. Of course the range over which those drones might operate from its controller craft might one day become quite a significant distance.



Of course machines have other issues and there would be a continual war of evolution of hacking and counter hacking measures. Both remote hacking and also physical. Last thing you want is someone to disable and then capture your top of the line drone and then turn it on you. It's also a public relations disaster if that AI controlled jet gets hacked and then turns its weapons on a built up urban area.

Indeed even once such instance could put the brakes on AI based war machines. It would fuel the fires and be bolstered by years of films which have depicted "rise of the machine" storylines and the like.



And alongside this we have to consider that not only is AI being worked on but also human augmentation. Exosuits are already being worked on and it wouldn't surprise me if we start to see the beginnings of a "Ghost in the Shell" type world.

Just to be clear I do not at all think we will see a future of autonomous, decision making, AI warplanes, I'm only talking about remote piloted vehicles.

To answer some of my own questions, Reuters has an good commentary on the subject.

https://www.reuters.com/article/us-military-airforce-commentary/commentary-drones-fighter-jets-and-the-future-u-s-air-force-idUSKCN0ZL0FH

 Overread wrote:
Of course machines have other issues and there would be a continual war of evolution of hacking and counter hacking measures. Both remote hacking and also physical. Last thing you want is someone to disable and then capture your top of the line drone and then turn it on you. It's also a public relations disaster if that AI controlled jet gets hacked and then turns its weapons on a built up urban area.

This is science fiction, nothing more. It's trivially easy to secure a system like this against outside hacking, and even easier to make it 100% impossible to use its weapons against designated "safe" areas.

This^.

Also, I thought I've seen some strategies that allows the F-22 (F-35 maybe?) act has the "motherships" of unmanned airplanes who basically acts as missle caddies for the warplane.

I would expect that manned aircraft become an anachronism at some point. Where that point may be, who knows.

Ultimately, several factors are involved in that. People have limitations, you can push an airframe's performance further without having to worry about a somewhat heavy and fragile "brain" that has to do things like sleep and are missed when they die. Drone aircraft can be much more disposable and suicidal, they can press attacks and sustain losses that manned aircraft cannot. Pilots are...expensive to acquire and train, and aircraft like the F-35 and F-22 are monstrously expensive and slow to produce, if we lost most of the F-22 fleet it would take a decade to replace those aircraft (if it's possible at all), losses are orders of magnitude less replaceable than something like WW2.

Humans will be directing these operations, but drone aircraft will inevitably take over. They can operate with greater performance envelopes, they can be made much cheaper and replaced much more quickly, will require less logistical support, and drain increasingly valued lives at a lower rate.

I think the answer to this question is easy to answer. The human being is the weakest part of an aircraft. It limits the aircrafts abilities, increases the demands on the aircraft design. Due to these facts I believe the answer is - once the technology is devolved/and put into circulation it will almost completely take over. How long will aircraft even be necessary though? I think that is a more interesting question.

Between railguns and lasers of the future - I'm not sure any expensive small craft will be feasible and is more likely to become a liability/ waste of resources. Though I think aircraft/space fighters will always be useful as scouts. A scout aircraft can be very very small if it doesn't have to carry heavy weapons.

I think we might definitely see more remotely operated aircraft. While Aircraft remain viable weapons anyway.

So we'll probably see a generation of aircraft which are remotely flown by human pilots. At least until AA technology advances to the point where anything that flies is DOA, which will shift all combat back to pure ground operations supported by long range observation craft. This will also likely make ICBMs obsolete as well(Nukes are useless if they will get shot down most of the time). Meaning we'll be back to WW2 style combat just minus air support. Tanks will be back to the fore as their natural predator(aircraft) will be extinct.

 Peregrine wrote:

This is science fiction, nothing more. It's trivially easy to secure a system like this against outside hacking, and even easier to make it 100% impossible to use its weapons against designated "safe" areas.

As a technician, let me take a moment to tell you you're full of it. Even stand alone systems can be hijacked, as the Iranians demonstrated not all that long ago by spoofing GPS and getting one of the US nice stealth drones to fly itself to Iran.

Oh, and that trick also means that what the drone thinks is a safe area and what actually is are two totally separate things.

You're a technician of what, exactly?

It's interesting this question focuses on fighters and other military airplanes. There are a lot fewer complications having AI fly commercial passenger planes. And if people won't trust an AI plane to fly them for a three hour flight to Vegas, do you think we'll trust an AI with missiles?


Automatically Appended Next Post:

 Kid_Kyoto wrote:
Just to be clear I do not at all think we will see a future of autonomous, decision making, AI warplanes, I'm only talking about remote piloted vehicles.

To answer some of my own questions, Reuters has an good commentary on the subject.

https://www.reuters.com/article/us-military-airforce-commentary/commentary-drones-fighter-jets-and-the-future-u-s-air-force-idUSKCN0ZL0FH

Remote piloted planes have lots of limitations that human piloted aircraft don't have. Visibility from drones is terrible, and there's a big delay in response time. These things aren't killer problems for the missions given to current drones, but for air superiority fighters they would be. Maybe those issues get reduced going forward, but they're not going away entirely.

The big advantage of AI planes is if you take the human out of the instant decision making entirely. A human might oversee the plane, but stuff that needs to be done very quickly, like ID of a target and firing, I think that has to be left to the AI for the concept to work at all.

 BaronIveagh wrote:
As a technician, let me take a moment to tell you you're full of it. Even stand alone systems can be hijacked, as the Iranians demonstrated not all that long ago by spoofing GPS and getting one of the US nice stealth drones to fly itself to Iran.

Oh, and that trick also means that what the drone thinks is a safe area and what actually is are two totally separate things.

As an engineer who has done software design for hardware control, you're wrong. It's trivially easy to secure a system like that because you can simply tell it to reject any signal it receives that doesn't exactly match the commands it is expecting, complete with encryption that is physically impossible to break faster than the drone system cycles through new keys. The drawback is that maintaining security of the system means limiting who has access to the controls (for example, troops on the ground that could be captured can't be given laptops with control software on them and access to operate the drones) and IIRC higher bandwith use for the secure links, but that's a price you're willing to pay for something as vital as armed drones.

As for GPS spoofing, it's also trivially easy to defeat. Install an inertial navigation system (slightly less accurate than GPS, but more than adequate for this purpose) and give it veto power of the GPS. If the GPS signal conflicts with the INS the drone shuts down its GPS navigation, and uses the INS to return to friendly territory for troubleshooting. The INS is 100% internal, with no interaction with the outside world that can be hijacked or jammed. You can potentially use GPS jamming/spoofing to mission-kill the drone and force it to disengage, but you can't turn it against its owners.

(This, by the way, is supposedly already in place on US drones, casting doubt on Iran's claims of how they got the drone.)


Automatically Appended Next Post:

 sebster wrote:
Remote piloted planes have lots of limitations that human piloted aircraft don't have. Visibility from drones is terrible, and there's a big delay in response time. These things aren't killer problems for the missions given to current drones, but for air superiority fighters they would be. Maybe those issues get reduced going forward, but they're not going away entirely.

Actually, air superiority fighers are the easiest drone problem to solve. Visibility doesn't matter when you have radar, networked sensors, and beyond visual range missiles. Throw a ton of missile platforms into the air, anything without approved IFF codes gets missiles thrown at it until it dies. In the kind of full-scale war where air superiority drones would ever be relevant civilian targets aren't going to exist, so you can assume that anything in the air without proper IFF codes is a hostile target that can be immediately shot down. Worst case scenario you shoot down a friendly drone that had an IFF glitch, but drones are expendable.

The hard problem is using drones for ground attack roles in support of friendly troops. There you're dealing with a fluid situation, targets that are much harder to identify and separate from allies, and human lives at stake instead of expendable hardware. That's where you need human oversight to make decisions.

 whembly wrote:
This^.

Also, I thought I've seen some strategies that allows the F-22 (F-35 maybe?) act has the "motherships" of unmanned airplanes who basically acts as missle caddies for the warplane.

The Reuters article mentions one program the Loyal Wingman where an F-22 has some unmanned F-16s flying with it. But I can't see the point.

Pilots already suffer from information overload and fatigue, how does adding other craft to their workload help? Sure 20 guys in a trailer a thousand miles away can run the drones better than one overworked pilot?

So that's the heart of my question, does someone on the scene have some advantage over a team in a remote location? Do meat eyes have some advantage that cameras do not? Because I'm not seeing it.

 Kid_Kyoto wrote:
Pilots already suffer from information overload and fatigue, how does adding other craft to their workload help? Sure 20 guys in a trailer a thousand miles away can run the drones better than one overworked pilot?

The primary advantage would probably be in stealth and jamming resistance. You don't have to broadcast long-range radar signals, you have the human pilot providing short-range control (possibly with line of sight systems using lasers, which are impossible to jam). And presumably you'd use the F-16s as expendable missile platforms, launching the initial salvo before the human pilot engages (if anything is still alive) and drawing the enemy missiles. If the F-16s survive the return fire long enough for pilot workload to be an issue as the human pilot enters the fight they could simply be ordered to return home.

 Peregrine wrote:
Actually, air superiority fighers are the easiest drone problem to solve. Visibility doesn't matter when you have radar, networked sensors, and beyond visual range missiles. Throw a ton of missile platforms into the air, anything without approved IFF codes gets missiles thrown at it until it dies. In the kind of full-scale war where air superiority drones would ever be relevant civilian targets aren't going to exist, so you can assume that anything in the air without proper IFF codes is a hostile target that can be immediately shot down. Worst case scenario you shoot down a friendly drone that had an IFF glitch, but drones are expendable.

The hard problem is using drones for ground attack roles in support of friendly troops. There you're dealing with a fluid situation, targets that are much harder to identify and separate from allies, and human lives at stake instead of expendable hardware. That's where you need human oversight to make decisions.

You didn't read my post. I said "remote piloted planes". That is, drones piloted from the ground. Like we have now. I focused my response on remote piloted planes because I replying to Kid_Kyoto, who said "I'm only talking about remote piloted vehicles". Your theory of a mass of drones operating as a network is cool to auto react to anything without the IFF marker is cool, but it's not really a response to my post.

 sebster wrote:
You didn't read my post. I said "remote piloted planes". That is, drones piloted from the ground. Like we have now. I focused my response on remote piloted planes because I replying to Kid_Kyoto, who said "I'm only talking about remote piloted vehicles". Your theory of a mass of drones operating as a network is cool to auto react to anything without the IFF marker is cool, but it's not really a response to my post.

The same principle is true even if you don't trust the drones with full autonomy. Air targets are easy to identify, easy to track, and easy to engage. You don't have the complex decisions like "is this a kid with a toy gun or a sniper about to shoot friendly troops" or "there are civilians within the blast radius, is it morally acceptable to engage this target". You aren't sitting there trying to interpret low-resolution camera images, hoping you can find some piece of information that confirms the identity of what you're looking at. Anything without the proper IFF signals is either confirmed to be hostile, or an expendable drone with no family back home to mourn its loss in a friendly fire incident. Whether or not the engage/ignore decision is made automatically or by a human operator on the ground it's still a very easy call to make.

(And, to clear up the potential misunderstanding, "remotely piloted" does not, for any drone, mean that the pilot is literally sitting at a computer with a joystick flying the plane in real time. The drone operator is doing things like saying "fly to waypoint X", and the drone executes the command. A drone air superiority fighter is not going to be like playing a flight sim video game with real hardware, where seconds matter and the operator is looking out the window to identify and pursue a target, it's going to consist of a drone operator giving shoot/don't shoot instructions and waypoints to patrol. And the engagement is almost certainly going to happen far beyond visual range, with 50-mile missile exchanges.)

 Peregrine wrote:
The same principle is true even if you don't trust the drones with full autonomy.

It becomes a nonsense point, because it becomes 'flying object identified lacking IFF' with no other information available. Whether the drone autofires or whether some human with no information other than that given by the drone fires is irrelevant, because the info supplied is so limited no human value judgement is possible.

Meanwhile, the bigger issue is that you've dragged this off in to a conversation about an entirely hypothetical future tech of a network of drones with amazing radar and very long range missiles. That is not what was being discussed, and the only way it can be part of the conversation is if it is the next, most definite step in air superiority. It's a cool idea but it certainly isn't the inevitable next step in air combat.

Here in the world we live in today state of the art fighters have human pilots who fly planes with a lot of capabilities, and while radar and stealth are very important, aerial performance still matters. Indeed, high-g maneuvers was one of the first reasons people started talking about taking the pilot out of the plane. But those maneuvers can't be done from the ground, the limited vision and time delay massively reduce performance.

 sebster wrote:
That is not what was being discussed, and the only way it can be part of the conversation is if it is the next, most definite step in air superiority.

Except my whole point is that it is the inevitable next step. The missiles are there, the sensor networking is there (and a major reason why the F-22 and F-35 are so effective compared to previous-generation fighters), and once you take the step of allowing armed drones for air combat there is no reason to use anything but a swarm of expendable missile carriers. There is no dogfighting ability that can defeat another 50 missiles being thrown at the target. There is no point in bothering with high-g maneuverability when you can, for a tiny fraction of the cost, have the drone send out a "shoot here" signal as it dies and buy a few more drones to spam missiles at its killer. The idea of a super-maneuverable "ace" drone is dead in a world of expendable missile platforms that can overwhelm it with shot after shot after shot until it dies.


Automatically Appended Next Post:
Now, the concept of BVR missile spam does lose value if you take the context out of a full-scale war where you can safely assume that any target without an approved IFF signal is a threat that can be freely engaged. In that hypothetical limited war you need manned (or at least well-supervised) aircraft that can make decisions, and engagement ranges are shortened by the need to confirm identification of a target before shooting. But in that case you aren't talking about a war between peer-level states, you're talking about more Iraq-style interventions in countries too weak to resist our gifts of democracy. And you don't need to invest in new drones to shoot down rusted out cold war era MiG-21s, the weapons we already have are more than adequate.

I think Airbus gets the future air power picture right with this.



Essentially manned aircraft take a back seat to drones for the actual shooting but are still there to control them and take over if necessary.

 Peregrine wrote:

As an engineer who has done software design for hardware control, you're wrong. It's trivially easy to secure a system like that because you can simply tell it to reject any signal it receives that doesn't exactly match the commands it is expecting, complete with encryption that is physically impossible to break faster than the drone system cycles through new keys.

And the issue with that is that if it's not an expected command, even if it's a genuine one, the plane does not respond. Further the check degrades overall performance and increases latency, which is already an issue when transmitting from Colorado to Baghdad in close support situations.

Also, remember that a well executed spoof gives the recipient exactly what they're expecting.

Students at University of Texas at Austin managed to replicate the Iranians feat, so I'll say that GPS spoofing would appears to work. Further, according to the Wall street Journal, insurgents were able to hack the video feed that goes back to the pilot in the US, though they were unable to alter the feed, they were able to watch, for $25 in commercial software and hardware.

 Alpharius wrote:
You're a technician of what, exactly?

IT specifically, which I've done for Governments and Casinos and banks and factories and PMCs ATM the US Federal government signs my paychecks. Though I don't work on drones, I've seen enough guys like Peregrine proclaim their software invincible to know the sort of trouble that's asking for.

 BaronIveagh wrote:
And the issue with that is that if it's not an expected command, even if it's a genuine one, the plane does not respond.

It really doesn't work that way. I'm not talking about an unexpected command in the sense of "I'm expecting to turn left here, this turn is to the right, REJECT". I'm talking about things like encrypting the control signal, and if a transmission it receives does not decrypt to a legitimate control instruction it is ignored. An enemy trying to hijack the drone can not duplicate the encryption, so their messages will decrypt to strings of random bits. A legitimate controller executing a change of plans will have the correct keys, their messages will be decrypted to legitimate control instructions, and the drone will respond to the new plan.

And then there's things like common hacking strategies not working at all. You can't download an .exe file into my hypothetical drone and tell it to run the code because no such interface exists to do it. It is simply impossible to give the drone a file or run code other than what is hardwired into its CPU. Saying "accept this file" simply gives it an invalid message, just like transmitting KLSDFGNJIOERJGKLJSDFGKLJERIOGJUIOWEJF()W$FJ()WEPMVL SWERIOJGI to it. It doesn't match a properly-formatted control instruction, it doesn't even make it through the receiver module.

Further the check degrades overall performance and increases latency, which is already an issue when transmitting from Colorado to Baghdad in close support situations.

Nope. CPU power is sufficient that the extra time required is imperceptible to a human. Spending 0.001 second longer on a task is not a relevant delay, and the check is purely internal within the drone's CPU. It's also something that the drone has to do anyway, to separate control instructions from random noise on the antenna.

Also, remember that a well executed spoof gives the recipient exactly what they're expecting.

And my point is that this is impossible for a properly-designed drone. INS is entirely internal, if you spoof the GPS sufficiently to get the drone to go somewhere it isn't supposed to go it triggers the "GPS outside X miles difference from INS" override, the drone shuts down its GPS receiver, and returns home using INS data only.

Students at University of Texas at Austin managed to replicate the Iranians feat, so I'll say that GPS spoofing would appears to work.

They did no such thing. They used GPS spoofing to hijack a cheap toy drone and a private yacht, neither of which would have the kind of security-focused design of an armed military drone. Obviously it is possible to spoof the GPS data and convince the GPS hardware that it is in a different position, the question is what happens next. A cheap toy drone trusts its GPS data 100% and has no backup plan because there is no need to worry about security on a $100 toy, and will respond to the spoofed signals without question. An armed military drone with onboard INS will not. Once the spoofed GPS signal diverges from the INS data (which can not be spoofed) it rejects the GPS data and switches to INS mode.

Further, according to the Wall street Journal, insurgents were able to hack the video feed that goes back to the pilot in the US, though they were unable to alter the feed, they were able to watch, for $25 in commercial software and hardware.

So what? The video feed is weakly secured because hacking it gains you very little, and is something that you want to make available to lots of people. Breaking the weak encryption on the video feed has nothing to do with breaking the security on the control systems.

 Peregrine wrote:


It really doesn't work that way. I'm not talking about an unexpected command in the sense of "I'm expecting to turn left here, this turn is to the right, REJECT". I'm talking about things like encrypting the control signal, and if a transmission it receives does not decrypt to a legitimate control instruction it is ignored.

That works great in the lab, but in the field signal noise can lead to the encryption turning into garage. Believe me, I tried to get a signal across a room full of lathes and milling machines once. EM interference can be a beast even at short ranges if it's intense enough.

 Peregrine wrote:

Nope. CPU power is sufficient that the extra time required is imperceptible to a human. Spending 0.001 second longer on a task is not a relevant delay, and the check is purely internal within the drone's CPU. It's also something that the drone has to do anyway, to separate control instructions from random noise on the antenna.

If that's all the chip is doing, you have a point, but remember that the more calculations you pile on, the more the system performance suffers. If someone is actively trying to flood the control signal with garbage, that 0.001 is going to start multiplying quickly.

 BaronIveagh wrote:
That works great in the lab, but in the field signal noise can lead to the encryption turning into garage. Believe me, I tried to get a signal across a room full of lathes and milling machines once. EM interference can be a beast even at short ranges if it's intense enough.

Ok? We already know that the control signals are encrypted, so I'm going to trust the fact that the US is already doing it over your claim that it isn't possible.

If that's all the chip is doing, you have a point, but remember that the more calculations you pile on, the more the system performance suffers. If someone is actively trying to flood the control signal with garbage, that 0.001 is going to start multiplying quickly.

It won't start multiplying quickly, because if you have multiple commands coming in the signals overlap and data becomes garbage. If you're at a point where the "flood" is creating enough of a delay to be perceptible to humans what you're actually talking about is jamming, throwing up a wall of noise on the control frequency so that nothing can get through. And that has nothing to do with CPU performance.

Also, again, decoding and interpreting the control signal is something that already has to be done. You already have to compare the incoming binary data against a list of acceptable messages, so that a corrupted message is properly rejected. You're never going to be blindly accepting incoming data and telling the CPU to run it, whatever it is. This is just basic software design for communication.

 Peregrine wrote:

Ok? We already know that the control signals are encrypted, so I'm going to trust the fact that the US is already doing it over your claim that it isn't possible.

I didn't say that it was impossible. I did get the signal across the factory floor with some work. What I'm saying is that you're failing to remember the most basic element of IT security: KISS. The more heavily encrypted, secured and locked away a system is, the more difficult it is to use and support, as well as hack. An example: the Social Security administration recently went from unencrypted desktops to encrypted laptops. The issue? Lose a BIOS chip and the whole thing is scrap along with all the data rather than just having to move the hard drive to a new laptop. The funny part? It was a massive downgrade in performance both on each individual machine and also system wide. And, personal opinion, it wasn't any more secure than the desktops, which had been kept in locked and secured facilities.

What you're basically saying is that you'd encrypt it with a form of rolling encryption that changes every so often along a pre-configured set of keys. There are practical limits on the complexity of the encryption due to signal loss, interference, jamming etc.

The two ways I can think of to beat this are A) capture an encryption device, or B) monitor the transmissions and use the brute force approach to try and decrypt. I seem to recall someone using zombies to break 128bit in under two weeks, so it's not impossible with enough computing power behind you.

 BaronIveagh wrote:
What you're basically saying is that you'd encrypt it with a form of rolling encryption that changes every so often along a pre-configured set of keys. There are practical limits on the complexity of the encryption due to signal loss, interference, jamming etc.

What I'm saying is this is exactly what is already being done. Your objections about the limits on the concept are irrelevant because they have already been overcome.

A) capture an encryption device

Impossible. The encryption codes are stored in two places: in a secure US facility, and in the drone itself. If you have access to the ground-side codes you have already infiltrated and compromised the entire drone program, and any plan that starts with "first, we create a catastrophic multi-level security failure in a highly protected military program" is something for fiction, not reality. If you have access to the drone you have nothing, because whatever encryption keys are stored in the drone (assuming you can even recover them) will be obsolete by the time you can extract them and attempt to use them for anything.

or B) monitor the transmissions and use the brute force approach to try and decrypt. I seem to recall someone using zombies to break 128bit in under two weeks, so it's not impossible with enough computing power behind you.

It is impossible with modern encryption (and 128 bits is laughably low). Brute force times are on the scale of "if you have every CPU in existence working on the problem you can maybe break it within a few million years", and that's for the consumer-level stuff like what I'm running on my PC as an anti-theft precaution. Even getting the time down to two weeks is nowhere near enough when the drone is rotating codes every 15 minutes. It is simply not possible to break modern encryption in any practical manner.

And no, this is not something that can ever change, it's a product of math. Adding another bit of encryption adds exponentially to the brute force time, but adds only a tiny delay in operation if you have the key. Going from 256 bit encryption to 512 bit encryption would increase the brute force time of a system that takes one minute at 256 bits to longer than the entire life of the universe at 512 bits, but the increase in the delay in decrypting a message using the stronger encryption would still probably be imperceptible to a human. This makes it trivially easy for encrypted systems to keep ahead of brute forcing attempts, simply by adding more bits to the key.

 Peregrine wrote:


It is impossible with modern encryption (and 128 bits is laughably low). Brute force times are on the scale of "if you have every CPU in existence working on the problem you can maybe break it within a few million years", and that's for the consumer-level stuff like what I'm running on my PC as an anti-theft precaution. Even getting the time down to two weeks is nowhere near enough when the drone is rotating codes every 15 minutes. It is simply not possible to break modern encryption in any practical manner.

And no, this is not something that can ever change, it's a product of math. Adding another bit of encryption adds exponentially to the brute force time, but adds only a tiny delay in operation if you have the key. Going from 256 bit encryption to 512 bit encryption would increase the brute force time of a system that takes one minute at 256 bits to longer than the entire life of the universe at 512 bits, but the increase in the delay in decrypting a message using the stronger encryption would still probably be imperceptible to a human. This makes it trivially easy for encrypted systems to keep ahead of brute forcing attempts, simply by adding more bits to the key.

And yet, the reason we couldn't export 128 bit encryption was that it was too much and it could never be beat before the heat death of the universe, it was impossible and exporting it would open it's use to all those foreign countries and terrorists.

It was beat fairly rapidly after that.


Oh, and while getting onto base at Edwards might be hard, how about hitting the US drone base at Diori Hamani International Airport in Niger? Or Kismayo Airport, Somalia (less than 100 men in the whole country, supposedly). Don't get me wrong, these places are defended, but are also in unstable regions or outright warzones. Several of those encryption devices almost fell into the hands of terrorists when a coup took over Mauritania. As far as getting one from a drone itself, the US loses about 150 drones across Predator/Reaper/Grey Eagle per year. however it can be assumed that various branches have differing encryption schemes.

 BaronIveagh wrote:
And yet, the reason we couldn't export 128 bit encryption was that it was too much and it could never be beat before the heat death of the universe, it was impossible and exporting it would open it's use to all those foreign countries and terrorists.

It was beat fairly rapidly after that.

Followed immediately by adopting higher bit counts. Seriously, this is not a problem that anyone is worrying about. It is common knowledge that as CPU power increases we will need to keep increasing the bit count to maintain "heat death of the universe" brute force times, and that is why we keep increasing those bit counts. The consumer-level " you, thieves" encryption on my hard drive is 512 bits, and would require 39000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 times more CPU cycles to break compared to 128 bit encryption. You are displaying a complete lack of understanding of how encryption and security work.

Oh, and while getting onto base at Edwards might be hard, how about hitting the US drone base at Diori Hamani International Airport in Niger? Or Kismayo Airport, Somalia (less than 100 men in the whole country, supposedly). Don't get me wrong, these places are defended, but are also in unstable regions or outright warzones.

Great. You hit the base. The US now knows that the drone has been compromised (since you just killed a bunch of drone operators and security guards and such), and now you have to get it off the ground (using a copy of the entire control system) before the US response shoots your assault force and turns off the drone. You're talking about a plan on the level of "break into the US air force base, climb into the cockpit of an F-22, and fly away with it". Best-case scenario you get a drone that you can't operate very effectively because you don't have the rest of the infrastructure, more likely you shoot some people and blow up the drone.

Several of those encryption devices almost fell into the hands of terrorists when a coup took over Mauritania.

So what? Getting the device gives you nothing because the keys contained with the device will be obsolete by the time you can extract them. The reason encryption is so effective is that even if you have the hardware that uses the encryption it is worthless without the actual keys, and keys are easily replaced. In fact, with a highly secure military program like armed drones, the keys are being replaced on a regular schedule and you might only have a few minutes to use a key before the next one takes over.

however it can be assumed that various branches have differing encryption schemes.

It can be assumed, but that assumption doesn't mean anything. The math behind encryption is public knowledge. The algorithm used to encrypt my hard disk has been published in scientific papers, and the encryption software itself is open-source. But that gets you absolutely nowhere, because the key itself exists in exactly two places: in my head, and temporarily in RAM while my PC is running. The exact same encryption scheme has survived government attack in major criminal cases, where the prosecution has been forced to admit defeat because even with the FBI/NSA/whatever attacking the files they can't get in.

My concern with the overdependance on drones is that ANY system that relies on wireless communication control can be jammed. Forget about hacking into the encrypted system, just jamming the frequencies that commands are sent from would render this thing a paperweight.