 Good morning, everyone. Thank you for making your way over to the Aviation Village on this early DEFCON morning. It's early for DEFCON, right? Oh, yeah, I know. It's a pleasure to be with you today. I'm Professor Andrea Matwishan. I teach in the law school and the engineering school at Penn State and run a policy lab called the Policy Innovation Lab of Tomorrow. And that stands for pilot also. So it is with great pleasure that I get the privilege of moderating this panel in the Aviation Village on aviation safety. So we have an esteemed group of panelists with you here today, who I will introduce in a minute. But first, I would like to take a minute to express my personal appreciation for the existence of the Aviation Village. It's something that I think was lacking from the DEFCON community. And I think it's something that was lacking from the broader social conversation as well. So I'm someone who works with members of Congress and regulators regularly. And I'm asked which industries are engaging with the security research community and which industries are not. And the level of engagement correlates directly to the perception of care that exists in regulatory circles. So it's great to see this village existing. It's great to see industry engaging with the security community. And it's something that regulators will notice. And then I will be sure to share with regulators because for better or worse, they ask my opinion regularly. So without further ado, let me give you just a bit of framing in two sentences and then turn it over to our esteemed panelists. So what we see as with every industry that is shifting to a more technologically driven set of models, we see a learning curve and challenges in design and insecurity in aviation. In terms of the policy initiatives, obviously the most recent set of incidents that have brought policy consideration into the fore have been the Boeing 737 MAX incidents. And on March 13th, of course, the FAA suspended operation of the 737 MAX. On May 16, Boeing announced the completion of a software update. And when we last left the evolution of the fix and improvement, on June 26, the FAA had identified one additional requirement according to the Boeing website. And so the evolution toward even better safety is on its way. And so we see this progression in this industry as in the medical device industry, another industry that is engaging aggressively with the security research community and regulators to work toward making the public safer. And this stands in contrast to some other industries that are not with us today at DEF CON. So with that, let's take a strolled-on memory lane in the case of some of our panelists. And I will introduce them now. So Stefan Savage, who will be up first, is a professor in the Department of Computer Science and Engineering at University of California San Diego. Then we will have Renderman, who is a security enthusiast with a focus on security threats of all sorts. He's the founder and principal researcher of the Internet of Dongs project. And we have, and last but not least, we have Spanky, who is, yeah, it's kind of not on your Twitter profile, so I don't actually know it. But you do, so you'll share it with us. So Spanky is a current CISO and retired Air Force fighter pilot. He's a fan of Infosec and cocktails and comedy and CrossFit, and his views are his own, as are mine always. And I'm sure that's true for all of us here on this panel. So without further ado, I'll turn it over to Stefan to get us started. And I hope that you will all engage robustly in Q&A and we'll be back to include you in a bit. Cool, thanks a lot. So I'm the Schmo at this thing with slides, because somehow I thought there were going to be slides. But so I'm going to talk a little bit about the experience that I and the group of us that have been working together have had working on aviation security and how it compares to other things. So just some personal background. About a decade ago, a bunch of us at the University of California, San Diego, and the Folkestone University of Washington went off and just bought some Chevy and Pallas. And we said, let's teach ourselves about cars. And we took those things apart, reverse engineered them. And over the course of about a year, we then showed up at GM and Department of Transportation and said, we can remote take over all the on-star equipped vehicles and turn off the brakes and so on and so forth. And that kind of led to a lot of changes. And then about four years ago, one of the people in my group, Kira Lepchenko, who is a total car, sorry, plane nut, said, why don't we get the band back together and look at planes? And so we did. In fact, we got a bunch of that team back together. And we started buying 737 parts. And recording every aviation signal in Southern California. And then started reverse engineering all this stuff, teaching ourselves how planes work. And so I'm going to talk a little bit about some of the similarities and differences and a little bit about planes so you kind of can engage in this discussion if it's not really familiar to you. So the first thing, very similar. Planes and cars, really complicated distributed system. There's several dozen different computers, primarily doing things like navigation and flight management, which are all about improving efficiency for flight. And then a variety of things that are there for safety. Every one of these in a modern aircraft frequently is a different computer, and they all have to talk to one another. This is just like a car, actually. This part is super similar. The architectures are actually somewhat different. So pre-1995, which is most of the planes that you fly today, have what are called a federated architecture. So every capability is in a separate computer written by a separate vendor. And there is no shared network. So you draw a line between, you draw a wire between units when they need to communicate. Something called the 429 bus. It's directional, so in fact, you get to send messages in one way. If you need a response, you drag a 429 bus in the other direction too. And so this is, again, most of the planes that you all flew if you flew to come here. Starting in 1995 with the 777, they moved to an integrated model where you do get a shared network. So 787 and A380 is basically an aviation version of Ethernet. It's not everything you think of with Ethernet. Redundant paths, there's no routing. There's control policing on bandwidth and so forth. But there is a shared network. And there's a shared substrate for computation. So instead of having lots of different pieces that do different things, you'll have basically blade servers where you can swap out parts and change the functionality. All right, so the attack surface, again, is somewhat similar to where you get external communications coming in. So these are things like signal demodulation for all of the ways that the aircraft communicates with the external world. These are going to be in things like the VDR. Then there's higher level protocol processing. ACARS, ADSB, RenderMan will talk a little bit about the ADSB signal. I think there's a variety of these high level protocols that get passed around. They are also interpreted by computers. And then on the ground, we effectively have Wi-Fi in some planes through something called gate links that allows them to do data loading once they get to the gate. And then when people actually have to update the flight management data loads or the software on these things, there are these little 615 connectors that you just plug this glorified Windows laptop into, and that's how you update the software. And all of those are the classic parts of the attack service. To a first approximation, in-flight entertainment and the cabin stuff is not a risk. Modulo, this talk that Ruben gave recently, about the 787, which is the only place one could begin to make such a claim. In general, there is no connectivity in most aircraft between those things. So one of the things that's super different, aviation is really regulated. So there is a really tight requirements around the process for every piece of software that gets generated, DO178C, and it requires them to certify the components in the environment in which they operate for particular airframes. It's slow, it's expensive, it means they spend a lot of time thinking about it. There are not until recently very much in the way of cybersecurity regulations, although it looks like we have some coming online in 2019. The consequence of this is that the software looks very different than what you find, even in cars and other embedded systems. Like, it's very common, there's no dynamic memory allocation because you can't count on what's gonna happen if new fails. So everything gets statically allocated. You don't have dynamic process creation. There are 10 processes, and they run for a certain amount of time. And they've done schedulability analysis to make sure that they all will run to completion in the time that they have. People do model checking. It is, the code quality is dramatically higher, I would say, from going through and looking at a bunch of this stuff. Now, the flip side is that because the time to certify is so high and the cost associated is so high, this stuff is ancient. The software is ancient, the hardware is ancient. When you fly a 737, most of that software was written in the 80s and 90s. So just to put a pin in this, if you fly on a 737, which is what I know the most, there's a unit called the Communication Management Unit, which is basically a thing that routes external signals to whichever other component needs to look at it. There are two different vendors. One of them is a 486. The other one's a Pentium, all right? And not like a Pentium II either, like a Pentium. And I don't know how some of you look like you are as old as I am. And on the Rockwell Collins Unit, the C++ code is compiled with a Wacom compiler, which you may remember if you're old enough to remember that vendor. And this stuff does not get changed because it got certified, all right? And so you would change it if you have to, but by and large, this is there. And so one thing to remember about this, it's very hard having a legacy thing that lasts when you have to anticipate, like integer overflow as a vulnerability was not a thing that people knew about when they even wrote this code. So finally, I want to close with some thoughts about how this is actually really hard work to do in practice. So it's really hard to get access to this stuff. No one has got a budget for a plane. And getting the parts is tough. And we've had weird, like having to get some weird connect in Thailand to ship us a part because they made it available. And it's, you know, we've been ripped off. And then getting documentation is, you know, you can't go to the vendor and get the documentation. Then making it work is tough, right? You know, like 400 Hertz power, which scares the crap out of me whenever we have to turn it on because that's what's in the airplane. You have to convince it, it's on an airplane. On the car, we would just, you know, record what was going on in the car and replay it to convince these things that they were in a car. You can't do that. I don't have a recording of what the hell is going on in an actual airplane. So you have to actually figure out what everything does. And then infer its impact. Finally, and this is, you know, these first two are boohoo, we managed these. The last one is really tough. And this goes to kind of the existing aviation security research and the challenges that it has had, which is there's really, it's really difficult to evaluate if you actually found something or not. Because you don't have, you don't have the plane. The DHS had an effort, which I, we can talk about offline, which attempted to do this was maybe not a huge success. And you also, in addition to figuring out how it works in the full on-varnished complexity of an airplane, hadn't you to understand how pilots think and what's just gonna happen with air traffic control? And so you need a lot of, like I feel a lot of times when we do this work we're part intelligence agency and like part just reversing, it's a very different experience. And so that's all I have for now. And we go on to the next speaker. Yeah. Good morning. I'm Render Man, I'm an alcoholic. Wrong meeting. So I don't have any fancy slides, but I find it interesting that you were talking about, budgets and access to things and everything. You have like school budget at least for research. I was an idiot on a couch. That's how I got into this. So six years ago, Def Con 20, that's seven years, six. It's the morning, don't ask me to do math. Yeah, insert math here. Give a talk on hacking ADS-B and potential other threats to aviation systems. That all started about a year or two earlier when I bought an app on my phone from Plain Finder. Plain Finder AR where you basically see a con trail in the sky, you look through the camera on your phone and it overlays the information about the flight. I'm like, oh, that's cool. Saw something online where it was pissing off DHS because they were like, oh my God, people know where, what flights are and everything like that. And it's like, okay, I gotta buy this quick before something gets cut off. Starting curious about, okay, where does it get this data? How does it know? Phone, I know only has a certain amount of RF receiving so it's like, it can't be like a Wi-Fi or Bluetooth or anything coming off the plane. So where is it getting this? So I started researching, I found ADS-B, automatic dependent surveillance broadcast. Basically, the planes now are moving to a system where they have GPS on board and are broadcasting out to the world to air traffic control and anyone else that wishes to listen. I'm like, okay, that's interesting. There's a bunch of enthusiasts that run ground stations that pick this up and then turn it into a database and do real-time maps of the air traffic in the world. It's really kind of a cool crowdsource effort. But then I started looking at the protocol itself and reading up on it, started with Wikipedia and went from there unauthenticated, unencrypted. These are two words that as a hacker scare the crap out of me or make me giddy as a schoolgirl just because depending on if you're trying to break something. Basically, the protocol was designed in the late 90s and well before SDRs became a thing, they thought that, oh, there's only like, an airport will be the only ones buying these and only Boeing or any others will, they'll cost millions of dollars. We'll know exactly who's buying them. You'll have basically security by obscurity to transmit on 1090 megahertz. Well, then SDRs come along and you can take a USB TV tuner and program called Dump 1090 that now you're listening to raw ADS-B directly from the plane. That changed their entire threat model and the problem is that the next gen aviation system that ADS-B is part of, they were rolling that out for 20 years. Their modernization stretched 20 years. I find that amusing. Realizing the implications of this because when you say unauthenticated, unencrypted, well, can I spoof a plane? Can I spoof the signal to plane? Get that into air traffic control systems to confuse, scare. Can I just use this for say, like targeting of something? If I know exactly where the plane is and it's giving me a regular beacon, drones are cheap now. A lot of them are starting to come with ADS-B receivers. The more I dug into this, the more I got scared. There had been other little bits of research out there, in the published world and just anecdotally, but nobody had done any sort of in-depth dive, I found, at least publicly, into how is this mitigated? There's references to multilateration, which is basically the time from the plane to, for the signal to multiple receivers. You can approximate, it's basically the inverse of triangulation. Figure out where the signal's coming from. You correlate that with primary radar, we're actually bouncing signal off of metal, and hopefully you're within this fuzzy zone. The problem is that in the same documents that are talking about these mitigations of multilateration, they're also saying, we wanna turn off primary radar everywhere because it's expensive. This didn't make a lot of sense to me. Nobody had any documentation or write-ups on how the ATC scopes and receivers, how they differentiated echoes, duplicate beacons from these aircraft. There was nothing saying how they had dealt with these edge cases. Every time I go down this path, I find, oh, trust us, or it's secret. Two more things that don't work well with a hacker. Again, all this is just stuff I'm finding in public documents, sending a few emails, not actually putting, I have no aviation background other than being meat and cattle class, so trying to pierce that veil was really quite difficult. The more I looked at it, the more scared I got. So I figured I had enough to say, I can't prove to beyond a so reasonable doubt that they've mitigated these things. Time to get somebody else to help pierce the veil, so Defcon 20, I gave a talk, hacking the unfriendly skies, where I basically, scientific method, you make your hypothesis. They've mitigated these potential vulnerabilities and you keep trying to disprove that. I couldn't, so I presented my evidence and said to the crowd, please, prove me wrong. I wanna come back here in a year and explain, yes, they have mitigated this, here's the thing. Hopefully building some bridges to the aviation industry. A friend of mine was telling me yesterday that he remembers after the talk, the scrum of people wanting to talk to me and the yelling. There were so many people there that had a vested interest. There were pilots or people in the aviation industry that were like, great talk, you're a full crap. I'm like, okay, please, tell me how. Have you got documentation? They're like, oh, but a pilot has ultimate control of the plane, they're gonna make the decisions. Okay, great. But the pilots make decisions based off of data. Data comes from the instruments. If I'm able to introduce data into the cockpit, that's bad. So essentially the discussion came down to put up or shut up. If you say I'm crazy, prove it, please. I'm more than happy to be crazy. It's nice in here. Six years later, or five years ago, November of 2017, as was mentioned, there was a DHS report where they actually got A-757 and a couple of days at an airport. And I believe the quote from the report was, within a day and a half, we're able to get non-coroperative interaction with flight control systems with stuff that they could legally bring on a plane. Final report is obviously classified, but reading between the lines and what little birds have told me, yeah, ADSB was one of those, that they've been able to inject data into the cockpit. And that should never even pause. That some jackass can inject 50 extra flights into the system, can cause confusion at an unopportuned moment. No idea, no documentation, no idea if any of the vetting they did on these systems, what their threat models were. Because the other interesting thing I found was that all the security stuff for ADSB, all the testing, everything, was given to safety people. Safety people think very differently than security people. They're looking for natural failures. They're looking for, do you have a backup? Can you tell it's failed? Check sums, redundancies, et cetera. Security is assuming everything is working fine as spec, but we're injecting bad data. It's got check sums. We're not tipping our hand that you're under attack, so you don't think to question the data. And trying to get that discussion with industry going was very difficult because here I am, a guy on my couch with an idea. You know, they're Boeing and Airbus and everybody else that are already tens of billions of dollars into this rollout and there's this fundamental flaw of it's unencrypted, unauthenticated. You can't back that out. So I plowed forward, but enough people have paid attention that at least in the back end, they're starting to look at things and say, okay, this is possible, maybe we need to work on our training, maybe we need to work on our processes to take this into account. I really think that the industry is waking up to the fact that they are now network systems. You've been doing the same things for so long in the industry, you're in this bubble. When you could start buying on eBay 737 parts and cobbled together enough of a working system to start having a person just on their bench understand this, that changes the dynamic. It's not you need many years of university and access to multi-million dollars worth of equipment. It's something that somebody can easily do on a hobbyist budget. Like I said, I didn't spend a dime on mine, on my work, it was just all public documents. But the industry, I think, needs to wake up and realize that if somebody's got an idea or a concern or something like that and is coming with evidence, a paperwork or something. Yeah, nine out of 10 of those might be, oh, you just don't understand the whole system, here's how it works. There's that one out of 10 that might have something. And might save your ass. And quite frankly, most of us flew here. I like my ass, I want to be able to get home. So I find it really ironic giving talks like this when you have to fly home. But yeah, so that's my story. I'll be teaching a bit about ADSB over in the village after this, and as long as I can stay awake. And thank you. All right, good morning, everybody. Glad to see you here on that happy note. Always setting me up for success, Render Man. So first off, let me start. My real name is Steve Lozinski. I do go by Spanky, I will answer any of those. And for those of you who came here specifically to hear Pete Cooper talk, I am the shorter, much more handsome replacement. So I apologize if you want to leave now, I totally understand. So I'm going to just build on what Pete was going to originally talk about, the idea of offering a pilot perspective from my experience. So like what Andrea said, I was in the Air Force 24 years, I had a great time. I flew F-15s and F-22s, thoroughly enjoyed that. And as a result of having such a good time, I spent my last three years at the Pentagon. So a bit of a repayment for all the places I was able to go and get into flying. But the good part of that, as bad as it was, and I have a hard time admitting to it, and the awesome, I might start dancing here in a second. I worked in Office of the Secretary of Defense, cyber policy. So the civilian side of the Pentagon, not something you would normally hear about, but the civilian runs the military, right? And representing military operations, things like that. So worked on a wide variety of things, aviation, cyber policy being one of those, helping with the initial charters to cooperate within the government to get after these aviation cyber problems. So I can offer you a pilot perspective from my time flying. I'm a CISO now, I thoroughly enjoy that. I'm not a hacker, I will hack the system and policies to get shit done. So I'll claim hacker in that sense. But I'm not gonna tell you anything about my street cred with the Wireshark class I took at B-Sides Vegas and tried my best on. So, but then also offer some government perspective. So I think exactly what Render Man was talking about. Industry has seen it, not necessarily talked about it and is now talking about it. Government's the same way. And when you look at government, industry, and the researchers, you all, the hackers, there's a whole lot going on with those stakeholders. And everybody wants to do the right thing for the flying public. That's our goal. That's the point. That's how we focus on what matters. So we quit worrying about that I had to help write a charter to get government agencies in the same government working on the same problem to cooperate with each other. That's our government. Okay, that's fine. It works, it's just tough to watch it when the sausage is being made. But otherwise, what I'd like to offer is, again, from that pilot perspective, what I think's important, and again, Render Man did a great job as my lead-in man for this, is the idea that I'm in a cockpit. It doesn't matter if it's me by myself and the previous airplanes that I flew working with the other guys in my formation. If I had a problem and how I solve it. Or if I'm in a commercial cockpit, which I've never done, but the ideas translate over, I'm talking to somebody next to me, I'm talking to folks on the ground, they're gonna help me solve the problem. I absolutely have redundancies and safety things we have learned over the years in aviation to build on that will help me take care of the problem. I really don't care what's causing it. It's nice to know that this fuel pump connects to this fuel line, and when I see that light, I know exactly what bolt fell off and is causing this problem. But when I'm flying and trying to take care of an aircraft, I'm either gonna hurry up and land, I'm gonna kinda hurry up and land, or I'm just not gonna worry about it that much. Whatever got me to that point. So I have to rely on the instruments, I have to rely on the data, and what Render Man was talking about, I think the difference there is we tend to get into the very highly exquisite, dangerous cybersecurity fun to watch in a movie because it's really exciting and to see how that happens and maybe the wing falls off the airplane because of a computer, that would be a great movie. The reality is it's the less exciting, I'm not gonna call it boring, it's the kind of work that Render Man does that if he's messing with basic navigation, I've got backups, awesome, I can handle that. Something happens with my radio and my communications, okay, that's fine, I've got backups for that. Put those two together, well okay, I can still handle that, I practice that, but it's a little bit harder. Now you make it where I'm in the weather and I can't see where I'm flying, that sucks, that's terrible, and oh by the way, I just lost a engine because of a fire and I'm now having a hard time flying the plane and comms and now that's the problem that I don't ever want to have to hear about or face because of the complexity there. So it's the small underlying things that are just as hard, that are absolutely as important, they're just not always the most exciting and most fun that as all those stakeholders are talking to each other, that's the other area that I think helps to focus and again with that pilot perspective. So with that, I will stop yammering on. Andrea if you want, we are happy to stand up here together to help with the microphone logistics and I'll turn it over to you. So assuming the dais here holds come on up and we'll take questions from the crowd, are there any questions for the panel? Are y'all just gonna walk home after this? Okay well I have a question, oh no, good, good, please, yes. We're gonna repeat the questions as we answer them so that everyone can hear. So the question was based on things that are in the news, things that Render Man talked about and you put all that together, what's the level of danger? There is something there, my opinion though is that when you look at the totality of what's going on, one thing by itself is bad but from again the pilot perspective there are multiple things built in, lots of redundancies, lots of backups that are practiced, I'm gonna say are taken for granted, it's just what we do in the cockpit. We may not recognize it but we have other backups that will help us get to recognizing that problem. So again it's not easy but there are things to handle it. The extra part of the answer I would offer is if my mom asked me, hey I heard this and I saw this and you were at this conference and that's terrible, I'm not getting on a plane, I'm like, mom I flew home, I'm not gonna walk, Render Man's not gonna walk, it's a long walk where we both live so, I think just in the big scheme, flying's dangerous, bad things happen, we all still fly, we learn in the flying community, we come back from that, we learn from accidents, we don't want it to happen, we'd like to get ahead of it, absolutely. So that's why having the village, having these talks, bringing people together, there are bad things, goodness is coming out of it and improvements are starting to take place. I just wanna add one thing, you're saying flying is dangerous and everything but it's still safer than driving or walking across the street in New York, you're still much more likely to make it home by flying. So it's one of those, yes there are threats, let's just keep them theoretical at this point, so. So just add one last thing, I think one real challenge that the independent research community has is that we don't know, so Ruben had this great talk, he got this binary, I believe everything that they found, Boeing says it doesn't work because of XYZ and Q and they're probably right, but they're the only ones in a position to evaluate that because we don't know. Show me. Right, but this challenge is a particularly tough one in aviation because there is no third party who can check this stuff and so we're in a fairly awkward phase where we're trying to figure out how to have that relationship with industry and have it be a constructive one. So I'll just tack on that there are lessons to be learned from other industries in the way that trust is engendered and built. So you have industry letting trusted researchers in, working with them and then creating a vouching basically through collaboration and a recognition through adequate transparency that there is a process in place and that everything is heading in good direction. So trust is something that is both real in terms of the code but also performative in terms of making the outside public have a sense of calm about the existence of the right people working on things and that there are both technological and human processes in place to ensure that things are moving forward as best as they can in light of technological and physical design constraints. Yes. So the question was, oh sorry, yeah. So the question is, I'm looking for the aviation, who are the aviation have to go and if we hit start, let's get to where they are if you guys think they're in good place. So the question is with reference to the medical device manufacturing industry, is there a parallel? How has that security ecosystem developed in collaboration with the industry and how does that work? Is that security ecosystem developed in collaboration with outside researchers and regulators and what is the trajectory that we might expect for a collaborative effort in the aviation space? So I think it's a great question and a great example. So the biohacking village in Def Con has been around for a while but it really hasn't been, except for the last few years that we've seen a significant presence of medical device manufacturers and also the FDA co-runs that village in part. So there is a workshop where they have exercises to break devices in a cooperative way and the companies come, bring their devices and say please break these devices, tell us what we're missing. And the other piece of this is that there was a significant change in the law that happened in 2015. I was the lawyer on that. So there was a security research exemption that was created under the Digital Millennium Copyright Act which specifically included medical devices which meant that researchers could examine the coded medical devices for the first time without fear of copyright suits. And so it was this combined effort of engagement from regulators, engagement from industry, engagement from the security research community in a collaborative way saying we've got some issues. The goal is ultimately patient safety. How do we make this happen? So this transformation has basically happened robustly since 2015. So in four years we may look at a completely different aviation village that is even better, even bigger and more collaborative in working toward passenger and social safety. So just there is one really big difference between medical devices and aviation. Actually aviation and almost everything else. Like for medical devices, when Kevin Fu and Yoshikono and the Barnaby Jack did their stuff, they had the goods. Like here we can make the thing do the thing anyone can look at it and tell that that's true. When we did our car work and showed up at GM or Chris and Charlie did it for the fiat anyone could look at that and say, yeah, there's a real issue here. We know we have to do something. With a plane, no one yet has the goods. We have pieces where I worked at this in a simulator. I did it on a piece. I did it on a thing. I think this would happen. And no one wants to have the goods because no one wants to do this test in the real world. And so it creates a different dynamic because there isn't this other forcing function. Yeah, I'd also point out that with the medical devices, the complexity level is, or is of magnitude more in a plane. You've got thousands and thousands of different systems working together, whereas a medical device, one thing does it very specifically. So it's like having thousands and thousands of pacemakers and defibrillators and other things around there. But the other thing, another comment is you don't necessarily have to have the goods. You don't actually have to have the full plane or anything. One thing that a lot of people I've talked to that didn't like my research were saying is, well, it was on a test bench. It was with a simulator. It was with, it wasn't with the real thing. When you do it on a real thing, then I'll believe it. It's like, well, I'm never getting a 757 all by myself. But there are test labs. There are places they do certifications, verifications of equipment and such that are for all intents and purposes for their, on their side of things, are the real world. Getting access to those would actually be very useful. If you show up with, hey, I've got a thing that I think will work. I've got data from my bench that says it will work. Get access to those test labs. Say, like, again, show me, prove it. It could be for the industry or the vendor's benefit to do so. Working with someone is a lot easier than trying to build more walls between them. I'll add a couple other things. I think a huge part of your question and where it's going to go is the willingness to have the conversations, the willingness to do things like the bio village and those types of efforts. Not speaking for industry, but a good example of what's being thought of. The report that Pete Cooper is actually over in the village, he authored and it was put out by the Atlantic Council. Great discussion, first one of its kind of what are all the different perspectives and when I say government, industry, security, researchers, how do they work together and how do they have agreement? So at least starting that dialogue, having it not just be something in the background and the government side, but getting out here, having a village and the willingness to talk about these things, to acknowledge that they're an issue and then when you look at the folks that came by the village and are supporting it with defense digital service, folks from the Air Force, folks from DHS and the idea of yes, this is goodness, let's keep it going. So that's half the battle and that's a big part of continuing this effort and then watching it grow and then trying to help it accelerate. I'll just throw in one little bit before we take the next question. We also have the Hewlett Foundation to thank for sponsoring this panel and so there's interest in the private sector foundation space and in the public at large, intensely on these issues. So there are many constituencies seeking to engage and who want to help and be reassured that we're moving in good directions in all respects in this industry. So he was basically asking that since a lot of these protocols are unencrypted and authenticated and there are sample ways now known to spoof these, is there any effort to add authentication or encryption or anything like that? Short answer is systems exist for that. Military uses an encrypted version of ADSB in theater but they, the problem is when you introduce encryption, how do you change keys out when you've got 25% of your fleet in the air at any given time? If something, you get a single bit flipped in the transmission, the entire message is garbled. Like it reduces the effectiveness and resiliency of the system at that point. As for authentication, you could still do primary radar for instance, make sure there's actually something coming from there as well as probably better vetting of the ATC towers to look, how do they deal with echoes? I've got documentation that says some of the stuff, if it sees like two of the 24 bit ICAO addresses in an ADSB single, the system just freaks the hell out because it was never designed with it. There's no contention protocol. There's a lot of pilots I've talked to where you get enough planes in the area for like some big event or anything. Stuff just stops working because there's no contention protocol. They're just stomping on each other in the RF spectrum. Yes, there is better systems, but to get everybody to play nice and roll the stuff out globally at the same time, I don't want that job. And just to add one thing, the other challenge I think it has is that it doesn't solve anybody's immediate problem. So if you look at the advent of crypto in internet protocols, for HTTP we added encryption because it actually solved the problem that at least we thought we had. For a lot of other things, whether it's DNS or BGP or what have you, it would have taken a lot of work to pull it off and get all the keys set up and distribution and so forth, but it wasn't enough of a problem to make it happen. And so this is the kind of thing where like someone would need to decide that no, we're gonna make this a requirement. And I don't think there's been, it hasn't been enough of a reason to force through all the complexity that Render's talking about. So the ADSB and a lot of these things are through the ICAO, the International Civil Aviation Organization. So you basically have all these countries trying to come to some agreement. The fact that any of this shit works at all is astounding to me when you've got rooms full of bureaucrats and people with different ideas and different interests actually reaching a consensus for a standard. So the fact that any of this works is amazing to me. Changing something out doesn't happen at the speed of the internet. It happens to speed a slower than government. This will be our last question, please. So, I'll take this if that's okay. There was a couple of questions in there. One was with TCAS used for collision avoidance versus ADSB, which is just reporting to ATC. So TCAS is plane to plane. So you're saying, yeah, it's broadcasting its position. You need to, you know, they get too close. The issue orders, you know, one goes down, one goes up. But the thing is there is no way to verify the source of that signal. Yes, the signal must be there, but it could be coming from anywhere. And reporting, like what it's saying is where it is can be anything. On a plane, you have very limited amounts of directionality to work with. That's the thing is, now we're getting in parsing technical details, we can speak into that. And what was the second part, sorry? Oh, the primary and secondary radar systems that they're wanting to turn these off. Not necessarily, they're just gonna shut down everything one day. Primary radar is expensive. Secondary radar basically is where you're bouncing a signal that interrogates a transponder on the plane and it broadcasts out a reply. They're expensive to maintain over and don't work over, you know, oceans, over like Gulf of Mexico. So, and there's other areas like mountainous terrain and that where you can't don't have the line of sight to bounce signal off of metal. If they can whittle down the amount that they need, it saves money, everybody's happy in that respect. Places like Anchorage where they trialed ADSB, you basically have to fly down this narrow valley between the mountains in order to land. You can't bounce a signal off of that. And if it's reporting is wrong, it's gonna scare the crap out of ATC. If suddenly you're flying down this valley and your TCAS system says, hey, there's something in front of you. You've got about 600 meters on either side to work with. I don't wanna be on that flight. So it's one of those, yeah, I may be completely wrong, but I wanna know I'm wrong because I really don't wanna be right. And with that, we will call it a day. Thank you very much for joining us. Thank you to the Aviation Village for letting us share this panel with you. Thank you to the Huda Foundation for co-sponsoring the panel, and happy flying.