The Cyber Trust Mark Debate
With Chris Romeo, Izar Tarandach, and Matt Coles
The Cyber Trust Mark, a new FCC program aimed at assuring the security of IoT devices is the topic of discussion today. We discuss various aspects of the Cyber Trust Mark, the history of similar initiatives like UL certification, and the challenges faced by consumers in determining the security of their devices.
About this episode
The Cyber Trust Mark, a new FCC program aimed at assuring the security of IoT devices is the topic of discussion today. We discuss various aspects of the Cyber Trust Mark, the history of similar initiatives like UL certification, and the challenges faced by consumers in determining the security of their devices. They also debate the merits and drawbacks of regulations like the EU's Cyber Resilience Act, the importance of secure-by-default design, and the limitations of relying solely on consumers or independent labs to ensure security. Throughout, they explore whether this new mark can genuinely make a difference or if it's just a rehash of old ideas.
Follow AI Security Table:
➜ Home: https://securitytable.ai/
➜ X: https://x.com/AISecurityTable
➜ LinkedIn: https://www.linkedin.com/company/ai-security-table/
➜ YouTube: https://www.youtube.com/@AISecurityTable
Transcript · 47 min conversation
0:09Chris Romeo Folks, welcome to another episode of the Security Table where we're already having madcap fun as we hit record. So we began with madcap fun and we're just going to continue right along with it. As always, I am joined by Izar Tarendas, who appears to be Out of focus when he was out of focus and lost and whatnot. But we'll get him, we'll get him back in focus. And Matt Coles, who is dressed for winter.
0:38Izar Tarandach No, he, he is ecstatic.
0:42Chris Romeo He's ecstatic. Spell that ecstatic.
0:47Matt Coles Estatic.
0:48Izar Tarandach E-c-static.
0:53Chris Romeo Yes, that's correct. I, I was gonna say ex-SQB.
0:59Matt Coles Tell me you have, tell me you have pets without telling me you have pets.
1:02Chris Romeo Yeah, that's true. That is true. Um, since we believe in operational security here, none of us have pets, or maybe we all have pets. You'll never know.
1:14Izar Tarandach Okay.
1:15Chris Romeo The topic of conversation here is something called the Cyber Trust Mark. I believe I've said that correctly. Cyber Trust Mark, which, I mean, it sounds very impressive. Like, if you have the Cyber Trust Mark, you are safe and secure.
1:37Matt Coles If you have the mark, you can trust cyber. That's what it is.
1:40Izar Tarandach You are marked for security.
1:43Chris Romeo Don't trust anything is what they used to say in the '60s, or don't trust anybody over 30, I think, or Oh, we've, we've moved on.
1:49Matt Coles It's now, it's now verify and trust.
1:51Chris Romeo Trust. Well, no, trust and verify was, has been around since 1980-something as a very popular term. But so what is the Cyber Trust Mark? Somebody lay out, since I just learned about it literally 15 minutes ago, but I tend to think I have known about it all along, which is, I, I, I always thought that it was our good friend Marc French, who Marc is the Cyber Trust Mark because we trust him.
2:20Matt Coles Um, so, so what is the Cyber Trust Market? Uh, so it is now officially an FCC program, uh, Federal Communications Commission program for, uh, giving information, ensuring, ensuring the security of IoT devices and providing information to customers, uh, that they can use to ensure that they are following security best practices and making use of what the, the, the manufacturer or supplier of the, of the product offers. So the Cyber Trust Mark will come with, uh, some sort of testing process and, and, uh, and a set of criteria and a, um, and a handy dandy thing that customers can scan so they can get information about the product they're using.
3:14Chris Romeo Uh, that'd be a QR code.
3:16Matt Coles It would be a QR code. It would be a QR code. And I, and we know how QR codes work.
3:20Chris Romeo So, um, no one's ever printed their own QR code sticker and stuck it on a box or, or made a drone QR code that people would scan during, uh, some, you know, laser light show. And who would go like laser light shows?
3:33Izar Tarandach Yeah. No. And, and, and clearly everybody knows how to look at the QR code and, and know if it's a valid one and a new one, right?
3:39Matt Coles Of course. Yeah. 'Cause it's, that's readily apparent to the, to the human eye.
3:43Chris Romeo But it turns out that this is not new. This is, as you guys were telling me before we started recording, this is just a rehash of UL, which UL stands for what? I forgot.
3:58Matt Coles Underwriter Labs. Underwriter Labs.
4:00Chris Romeo Yeah. Cause I remember years ago when they started on this journey that they were going to certify IoT devices. I want, it's gotta be 10+ years ago when that whole thing kicked off.
4:12Matt Coles It's about that.
4:13Chris Romeo Yeah.
4:14Matt Coles Well, when it started, yeah. But not when it went into, into effect.
4:17Chris Romeo You're thinking UL?
4:17Matt Coles Did it ever hit? You're thinking UL? The UL.
4:19Chris Romeo Yeah.
4:19Matt Coles UL 2900, uh, was present for some period of time, got some traction, but, uh, uh, I think it's, I, I don't know for sure, but I don't think it's got, uh, that's, I don't know if it's still continuing or not. That's a really good question. We should ask, uh, somebody or something.
4:38Chris Romeo Well, there's, there's much bigger challenges to explore in the US Cyber Trust Mark. So where do you wanna go first? Should we trust Mark?
4:50Matt Coles I don't know.
4:51Chris Romeo Oh, Mark French, I trust. Uh, a Cyber Trust Mark based on the criteria that I've seen so far, I don't know that it makes a big difference in my world.
5:01Matt Coles Well, so wait, go ahead. Sorry.
5:03Izar Tarandach If it doesn't make a, a big difference on your world, Will it make any difference on the, on the world of consumers who are the, the, the target of this stuff?
5:15Chris Romeo Wait, I'm not a consumer.
5:18Izar Tarandach You are an educated consumer.
5:19Matt Coles Yeah. You're, you're, you're in the, you're in the upper echelon there of know and care, meaning K-N-O-W, not N-O. Yeah.
5:28Chris Romeo You don't care. Don't care.
5:29Matt Coles Good point. Right.
5:31Chris Romeo Yeah.
5:31Matt Coles No, that's good.
5:32Chris Romeo It's a good point.
5:32Matt Coles It's, it's just that most consumers don't know and don't know enough to care.
5:36Chris Romeo It's good that we recognize that in this conversation, right? Like we're actually not the target people of this because you think of the average person who's buying a wireless device, say they're not a tech person and say they don't really have a lot, they're not good with, they would self-describe as, I'm not really good with computers. They don't have any ability to choose a product. I mean, I shouldn't say that. They could randomly choose a product off the shelf, but they don't have any way to differentiate between the security goodness of product A versus product B, whereas we do, because we would look at all the documentation, we'd read Reddit posts about the security stance.
6:25Matt Coles We'd watch, you know, our favorite YouTuber product review looking at this, right?
6:30Chris Romeo Yes, we'd look at multiples of them and compare them and see, you know, is it—
6:33Matt Coles but that user, that user base that you're talking about, it should be relatively easy for them to go out to look at product A and product B side by side and go, this one has a trust mark, this one has a recognition from the FCC and this one does not, and go, I will preface the one that has, you know, prefer the one that has the recognition for better or for worse, right?
6:53Izar Tarandach But, but look, look at it this way. So what's the. The label is giving us, right? It's a QR code, which you're supposed to scan, and then it's going to give you information about secure configuration of this thing. So now I'm Joe Consumer. I go into Best Buy. I pick up two different products. I scan both of them. What's going to influence my choice here?
7:23Chris Romeo Well, I mean, do they both have the mark?
7:26Izar Tarandach They both have the mark 'cause they had to scan both of them.
7:28Matt Coles Yeah.
7:29Chris Romeo So then they both have complied with this set of criteria. And so I think in the general consumer's world, they're equivalent.
7:37Izar Tarandach No, but what, what does the, the, the mark, scanning the mark take you to? Right? It takes you to, I'm guessing just a list of pro—
7:47Chris Romeo like a, a certificate that says it was scanned. That's how we've done it in every other certification scheme.
7:52Izar Tarandach But that's, that, that's the differentiation here. It takes you to a page with information that has to be easy to understand. Lord knows what that means. And it's, it should include how to change the default password, how to configure the device securely, whether updates and patches are automatic, and if not, how consumers can access them. The product's minimum support period end of date. So if I'm looking at both of those, in my mind as a consumer, non-educated one, I'm gonna go with, ah, this one looks better and it looks easier.
8:27Chris Romeo Yeah. And, and that's really not solving for anything, right? No. Because the person who's out there that's gonna take this device, they shouldn't do anything to it. Like, no, the, the, the mark should result in a product that comes online that is secure by default. I got an idea. Why don't we create a pledge and have a bunch of people sign it and then they could just kind of self-validate by signing this pledge that they're gonna do all the right things. Wait, wait, wait, wait, wait, wait. Better at them.
9:01Izar Tarandach That's, that's revolutionary. Could we get a government department to stand behind it?
9:08Chris Romeo Maybe a four-letter agency. Maybe. Maybe even better.
9:11Izar Tarandach Maybe, maybe.
9:12Chris Romeo I think it will happen.
9:14Matt Coles Maybe one that has security in the name.
9:16Izar Tarandach But no, but we have to be—
9:17Chris Romeo Yes, security.
9:18Izar Tarandach No, no, we have to be honest because a bunch of know-it-alls are going to throw shade on it and we are going to hate it.
9:27Matt Coles Wait, wait, is that us?
9:29Chris Romeo I don't know.
9:29Izar Tarandach And months after that, and months after that, in a puff of smoke, we are going to get notices of CVEs and breaches with companies that actually signed it. Yeah.
9:40Chris Romeo Like, I don't know. Could that really happen in the real world?
9:43Matt Coles Could it?
9:44Izar Tarandach Could it ever happen?
9:44Matt Coles No, it could never happen.
9:46Izar Tarandach I think that it's one of those cases.
9:48Chris Romeo Happened to Jobs, by the way.
9:49Izar Tarandach Oh yeah, it did. We'll give you more about it later. But, uh, it's one of those cases where you're going to have companies looking at us and going, did I do that? But, uh, yeah, it's traumatic. But on the other hand, on the other hand, okay, so what I'm trying to get at is that Chris, you said that, oh, the QR code should take us to all kinds of certificates and certifications. That's what we've done before.
10:14Matt Coles Awesome.
10:14Izar Tarandach So they're trying to do something different here. My problem here is that all this different stuff that they're trying to do, all of it is spoofable.
10:25Chris Romeo Well, and the challenge is they're putting the onus on the consumer, which we know the consumer is not going to go through 57 difficult steps of configuring the device. Secure by default means give me the device out of the box, bring it up in its default state. And then if I choose that I want to go configure it and loosen it in some way, that's my choice as the consumer. Now, personally, I don't think you should ever even give them that option, but that's just me.
10:52Izar Tarandach And that's a great point because if you look again at what's coming in the documentation, how to change the default password, how about you can't configure this thing without changing the default password?
11:01Chris Romeo That's how we solve default passwords.
11:04Izar Tarandach How to configure the device? How to configure the device securely? How about not giving the option of it not being configured securely?
11:11Chris Romeo Yeah. I mean, I was at Cisco in the days of Cisco 123. And when there was a mass development effort, I mean, this is going back a ways.
11:21Matt Coles Hey, that is, that is 8 characters.
11:23Chris Romeo And it's, I just want to say, I just want to say I wasn't Yeah, but there, but there was, but there was a mass development effort that went, that we went through to remove default passwords from the OS.
11:35Izar Tarandach Yeah. And now it's Cisco123!! But you had to change it.
11:39Chris Romeo You had to change it.
11:40Matt Coles Capital C. Capital C. You had to change it though.
11:42Chris Romeo We made it part of the bootup, the installation process where you had to then choose your password.
11:47Izar Tarandach And it changes every 90 days.
11:49Chris Romeo But we took out the, we took out the default password in it so you didn't, So if you chose to put in a bad password, that was your choice as a user, administrator of the device. It didn't come with it. So we took out the ability for you to have it. But what this Mark is talking about though is they're still going to have the default password, but you just have to tell people how to change it. Like that's ridiculous. Why would we even bother?
12:14Matt Coles Think about the default. First off, think about the type of devices that we're talking about here, right? These are mass-produced, mass-market devices. Potentially, right, they're tens of dollars, hundreds, maybe hundreds of dollars, depending, right? Um, and, and some of them, it's probably, there are hard to use user interfaces, right? You may have a mobile app that is used to configure the system. And so it has to boot into a config, and it has to boot in such a way, you're not going to boot it right onto a network with a UI. Right? So let's think about, you know, if you think of like a, a smart speaker or, um, you know, a, a, I guess even a router to some extent, right? You plug that thing in, it has no UI. You have to connect to it somehow, right? So if it doesn't have a default password, if it doesn't, and I'm not careful here, I'm not arguing that this is a bad or good thing.
13:09Chris Romeo You, I think you're arguing for default passwords is what I'm arguing.
13:11Matt Coles No, I'm not. I, I'm just suggesting that there are some challenges with consumer devices.
13:16Chris Romeo But you know how most people have taken care of it? Because I did it in multiple devices that I've installed in my house in the last couple of years. They use Bluetooth from the mobile app to temporarily set the thing up in a— so you can, we can argue the security nature slash architecture of that solution, but it's a proximity problem for an attacker. Like, the attacker's got to be within 500 feet of my house when I'm installing the device.
13:44Matt Coles So I get some value or trigger a power failure so they can, you know, be driving by while, while your device is resetting. Yeah.
13:51Chris Romeo Right, exactly. Even then, once it, but once it's set up, it goes into a mode where it'll just reboot, right? Like you couldn't even drive by and turn the power off and back onto my house and have the device. It doesn't go back into learning setup mode. Once it's set up that first time, it starts to, it goes into its normal, you can hard reset it by pushing a button on the side for 30 seconds.
14:09Matt Coles Yeah.
14:10Chris Romeo And then it would come back in that state.
14:12Matt Coles There are ways to design these products, right? There are ways to design these products such that it makes it easy to do what we're talking about here. But these require, this requires effort by develop, by the designers. Um, and so I, so for a moment, actually, I want to, want to bring this back a little bit. Let's separate the Cyber Trust Mark, the, the label from the QR code. If the QR code is just an information delivery system, for here's information for the device and the Cyber Trust Mark is what actually customers should, consumers should be looking at to say, this went through some sort of validation process against a set of rules or a set of requirements, then, then it may be a little bit better. Devil's advocate here. Um, but, uh, but I 100% agree with Isar and both of you actually, that the rule should not have a default password. It should have a secure setup process.
15:09Chris Romeo Right.
15:11Izar Tarandach Exactly. And, and I think that the way that you put it, it becomes like the, the label that you have on, on cigarettes, right? Yeah. There's plenty of testing behind it. There's plenty of, uh, science behind it and people still keep smoking.
15:24Matt Coles Well, even a, even a UL mark, let's, I mean, if you look at something like the, the Cyber Resiliency Act over in the EU, right? It's tied to their, to their, their regimen, right? And so when you see the mark, It's not UL, but I forget what the specific is.
15:38Chris Romeo Yeah.
15:39Matt Coles Um, but when you see that mark on the device, you know it's gone through some eval— some validation.
15:44Izar Tarandach So that, that's what I wanted to talk about next. So we talked about the Cyber Trust Mark French, and now let's talk about the French Cyber Mark Trust, which is the Cyber Resilience Act, which is what they, they're doing in Europe right now.
15:58Matt Coles Right.
15:59Izar Tarandach And I, I think that Matt, you, you, you touch exactly on it. They are doing like mandating actual secure by design.
16:07Matt Coles Oh yeah.
16:08Izar Tarandach Actual things that people should do before their products go out.
16:12Matt Coles With financial penalties to the organization if you don't. Exactly.
16:15Chris Romeo We, well, we all know how good the government is at innovation.
16:18Izar Tarandach Well, it depends on which, no, but this is not an innovation thing. This is things that we know.
16:23Chris Romeo It's a squasher of innovation.
16:25Izar Tarandach No, no, it's not.
16:26Matt Coles No, it's not.
16:26Izar Tarandach No, no, no. I'm going to disagree with you here. These are the things that we have been telling people for how many years? And now the government is saying, hey, you know what?
16:34Chris Romeo Those guys are right. But when you— so introducing regulation, you're never gonna convince me that regulation creates innovation. It just doesn't.
16:44Izar Tarandach That's the thing.
16:44Chris Romeo Governments of the world do not create innovation. And if you have files—
16:51Izar Tarandach but we are not talking about innovation here.
16:54Matt Coles We are talking about their minimum best practices.
16:56Izar Tarandach We, we are both— no, we are talking the, the innovation happens in the kernel of the thing. And to the Chompy bits that give you the service. But we have been for a long time saying, people, there's a number of basic security hygiene things that we need to do, we need you to do, no matter what you're developing. If it's the most innovative thing in the world, if it's the oldest thing in the world, we still need these security things to happen. So now the government is coming and saying, you know what, those guys are right, these things should happen. No matter what you're developing. So this doesn't hit innovation in any way, shape, or form.
17:34Chris Romeo And it's never going to convince me that regulation is an enabler of growth, of product development, of anything.
17:43Matt Coles What you're suggesting, Chris, though, is that if there is regulation, if there's no— so if, if we can't get— if companies don't do security on their own, meaning as part of innovation, they don't build these capabilities in, and the only way to solve that is to create regulation, then, and that will stop innovation. There's a, there's a very big contradiction there. You can never have security is what you're suggesting.
18:10Chris Romeo So let's, let's see, let's look at a case study here. Let's look at an example.
18:14Izar Tarandach Okay.
18:14Chris Romeo Where do most tech products begin their lives?
18:20Izar Tarandach Taiwan.
18:20Matt Coles China.
18:22Chris Romeo From a company perspective, not from a manufacturing perspective, from a company perspective.
18:27Izar Tarandach Oh, you said that right.
18:29Chris Romeo They start, so startup companies are what is the innovation engine of the technology world. You can say big companies are innovation creators, but they're, they're acquirers of innovation. Sure. Most big companies don't, don't, I mean, you can give me a, you can give me a couple of examples of—
18:46Matt Coles Well, how, how big a company?
18:47Chris Romeo Innovation.
18:47Matt Coles What? When you say small companies, I mean, a couple hundred employees, a couple billion dollars.
18:53Chris Romeo I mean, start most, if you pick any popular consumer product that's in the market today and they started as a 1 to 5 person startup company.
19:05Matt Coles Sure.
19:05Chris Romeo But 10 plus years ago, Apple.
19:07Matt Coles Okay. If you take Apple as an example, right?
19:10Chris Romeo I mean, Apple did start with Wozniak and Jobs in the garage, right? Like that, the other guy who sold his shares for $100.
19:17Matt Coles What you're suggesting is every, every company is never, there's no company that is born into the, you know, massive giant size company.
19:27Chris Romeo And regulation is when you're a 5-person company trying to build something and someone tells you, here's this regulatory act that you have to comply with and the penalties are billions of dollars, which it probably wouldn't be for startups, but it's—
19:41Matt Coles CRA specifically is some percentage of your revenue.
19:44Chris Romeo Percentage of revenue, right? But if, so the age-old problem that we've battled in product companies, all three of us have battled this, has been product feature innovation versus security hygiene. And there's always, in the early days of a startup, as much as it pains me to say, there's always gotta be a balance. Because if you come in and say, we are building, we are a security-first company and how we're building this new product, you will more than likely not get to market in a timely fashion to the point where you will more than likely not survive and not— so there has to be a balance. And so my whole challenge with regulation is you're putting potentially a draconian set of rules, depends on how you interpret the rules, on the innovation engine of our world. Now, if you told me that Cyber Resilience Act was a scalable thing that allowed me to build things and grow my maturity as I find product market fit inside of something that I'm building, meaning grow my security. Like I don't have to have all my, everything figured out. Like there's a level 1, level 2, and level 3. And level 1, if you're under $1 million in revenue, you gotta be a level 1, but it doesn't require your own embedded PKI inside of your solution because that's, could be a bad example.
21:11Matt Coles I mean, I'm trying to think of one example of what I could do.
21:13Izar Tarandach Chris, Chris, you know what? I, I, even if you are right, and I believe that you are right, I'm going to say that I don't care. Because, no, no, simply, it's very simple. Perhaps it is time. For us collectively to take a breather and say, we've been running a lot towards this market thing and see where we go. Information has almost no value now because people just go and say, another bridge, whatever. Right? So, yeah, perhaps it's time to take stock and say, are we doing the right thing here?
21:52Chris Romeo And my dream was always that the market would weed out the bottom feeders. But unfortunately, I mean, to some degree it has, right?
22:03Matt Coles At the enterprise level, yes.
22:04Chris Romeo But not at the consumer level. But even like, think about like doorbells, like smart doorbells that I can buy now. There's one company that owns a big chunk of that market right now, and they're actually part of Amazon now. They were acquired by Amazon, Ring. And you can look at their architecture and you can, and they have a very solid approach to security. They're, you don't see them in the breach of the day type of news of things like, you know, but that's because it's their core business.
22:32Izar Tarandach Nobody's going to buy a lock to their houses that doesn't offer that kind of, of security, even if they are not able to understand it.
22:40Chris Romeo So then we don't need a Cybermark if, if the consumers are that educated about smart door locks.
22:46Izar Tarandach No, no, no, no. But that's the point. That's the point. That, that's the expectation. If I'm going to Best Buy now and I'm buying, uh, I don't know, a sprinkler controller. Okay, dude, what, what, what you're going to do? You're going to put more water in my plants or less water in my plants? Yeah, I don't care about security. I don't care about this. I don't care about the other one.
23:10Chris Romeo Right. I think that's the general state of the general consumer doesn't see we, we have an unfair advantage, right? And that we, we threat model for a living. And so, but when you, we see the threats in our, you and I, all of us would be like, somebody could DDoS this water controller and flood my whole backyard.
23:30Izar Tarandach No, but wait, wait, that, that, that's the thing. That's the thing. That's the thing. It's not that we see the threats, we see the impacts.
23:38Matt Coles Well, that's, it's a natural thing. You know what the threats are, you know what the impacts are, right?
23:42Chris Romeo No, no, no, no, no, no.
23:42Izar Tarandach It's a different thing. Because the consumer out there could be more educated about the possible impacts, even if they don't know what the threats are. The threats are a completely technical thing, but a consumer understanding what a piece of hardware that they bring inside their hall and their homes could possibly do to, to their security and their privacy. That's a different kind of education. And that's something that we have never done before.
24:06Chris Romeo But my whole point is I want that smart sprinkler. To just be secure by default out of the box.
24:12Matt Coles But you, but you want that to be, you want the organization, you want the companies that create that to be self-policing, which is the way it's been for—
24:20Chris Romeo I want the market to police it ultimately. Well, that's the free market economy though.
24:25Izar Tarandach But no, but, but that's, but, but again, that, that goes to the consumer because they move the market. If the consumer is not educated into what could possibly go wrong, right? Or, or how bad it is when something goes wrong, then they are not going to influence the market and that's not going to force companies to sell.
24:43Chris Romeo But now, but now you want the government to put their hand into it and, and, and what level, what, what levers do you have?
24:51Matt Coles I mean, so, you know, and, and again, right or wrong here, right? I'm actually, again, I'm not, I don't 100% agree. There, there is a, I think there is a place for regulation, but as long as it's not a stifler of innovation, but it's a supporter of innovation. And I think that they, they can go hand in hand personally. There is, there is a good, I think there is some precedence for that, but think about the Ring case that you just talked about, right? So a single startup that was Ring before Amazon bought it. If consumers, if, if some savvy consumers decided that, oh, this thing is detrimental to my my physical and, and, you know, and data health, whatever, you know, I had leaked privacy and, and allowed, you know, attackers to, you know, walk into my door without, without getting stopped by the doorbell ring or the camera. Not that that's happened or anything. That, you know, as a consumer, you'd have some leverage, right? People say, well, you know, you're, you're a 2 to 10 person startup and this product is just getting You know, gain market share. I'm going to go buy company X and not yours, right? And you have influence. As consolidation occurs and now Amazon, Amazon can take a loss on these things and consumers have less capability to influence the market. So what drivers do you have? What levers are there to influence companies to do the right thing from a security standpoint?
26:20Chris Romeo I mean, if the, if. If the consumers in the market drove the bus here and actually said, and actually bought products, didn't buy crappy products, then the product, it would fix itself even for a company as big as Amazon. Because believe me, Amazon owns Ring and Ring has growth targets for this year that are tied to people buying more stuff. Installing more stuff in their house. That's how business fundamentally works.
26:52Matt Coles That's how capitalism and this particular market works, right? Mm-hmm. Again, right or wrong, I'm not going to judge them going here, right? If they didn't have quarterly targets, would it matter?
27:06Chris Romeo Right?
27:07Matt Coles So, uh, if you were just pushing a product and you could take a loss, you could take a beating on it, right? And you didn't care because you had something else propping up your business, right? And Amazon already has AWS that probably brings in more money than—
27:20Chris Romeo but they don't want to, I mean, they don't, business doesn't exist to lose money. So they're not going to take a, they're not going to let their, they're not going to drop their security guard down and build something that's, they also have a reputation attached to their, people buy, buy, use AWS because they trust the security controls and properties that they think they get. With AWS, right?
27:44Matt Coles So, so that's a particular, that's a particular set of consumers that do that, right? So the consumer who's buying the $10 device out of, from Best Buy or from Costco or whatever, isn't the one with leverage unless you're talking about the entire potential customer base.
28:02Izar Tarandach You, you are not going to get to a situation where you're going to open Amazon on a certain product and on the reviews you're going to say one star. They don't initialize their IV on their keys.
28:13Chris Romeo That's what we're saying. We might. I wish we did, but I wish we would get there.
28:18Izar Tarandach But that's not the—
28:18Chris Romeo coming back to the general consumer though, right? That's the general consumer. They, the general consumer just wants something that works.
28:25Izar Tarandach Matt is thinking about the reveal that he wants to say.
28:29Chris Romeo Nobody will understand it. AI will just blot it out.
28:33Matt Coles No, and you know, you know, I'll get a call from their, their incident response team or their, or their vulnerability response team. Like you could have reached out to us on Twitter.
28:44Chris Romeo Security at said company.
28:48Matt Coles We were not going to solve that.
28:51Chris Romeo We're at an impasse. We're not going to, we're going to have to agree to disagree, which.
28:55Izar Tarandach No, no, no, no. I don't think that we are so far. I mean, do you agree that the FDA is saying don't use yellow 360 because it causes cancer on food? It's a good thing?
29:09Chris Romeo I don't even know what that is. So pick another.
29:11Matt Coles Red number 5, right?
29:13Izar Tarandach Yeah.
29:14Matt Coles Some dye that has proven to cause cancer in humans. If the FDA says don't use that, either, either, either don't use it or you really shouldn't use it.
29:24Chris Romeo But we're, we're crossing into another.
29:26Matt Coles It's regulation.
29:28Chris Romeo It's regulation. It's, it's medical in nature.
29:32Matt Coles Oh, hold on.
29:33Chris Romeo Hold on.
29:34Izar Tarandach Oh, hold on. Somebody pinch Adam. Somebody pinch Adam.
29:36Chris Romeo Hold on.
29:36Izar Tarandach Timeout. So.
29:38Matt Coles Do we remember the, do we, do we remember the concept of guardrails?
29:43Izar Tarandach Yeah, I fell off of them.
29:44Matt Coles Guardrails and paved roads, right? And regulation could either be a guardrail or a bump or a, either a guardrail or it could be one of those Jersey barrier things that force you into a lane, right? Mm-hmm. So we want the guardrails. We want the limits so that people don't, you know, they don't inadvertently go ahead and put something in a product that causes mass, you know, mass death and destruction or privacy loss and data loss. But we don't want to stifle the innovation. We want to allow them to allow companies to bounce off of those guardrails without getting seriously, you know, run off a cliff. You don't want the Jersey Barrier type regulation. It's what you're, what you said, which is what Most regulation is, which, right? So there's a balancing act, right? We want the guardrail regulation. We don't want the Jersey barrier regulation.
30:37Chris Romeo So now we want the government to balance something.
30:41Matt Coles No, we want flexibility. What we want is regulation to be, to be those guardrails.
30:46Chris Romeo Oh, we know, we know the government is all about flexibility. I always send them an email on April 14th and say, hey, I'm going to send you my taxes next year. And they never respond.
30:55Matt Coles You know, you take it to a, you take it to the street.
30:57Izar Tarandach That's because you didn't send it last year.
31:00Chris Romeo I should have sent it 100 years ago.
31:03Matt Coles I know we're gonna get a ton of hate mail on this, on this episode, but, um, oh, I know that we're going to have at least one person in a treadmill out there screaming at Craig Saldaña.
31:14Chris Romeo Okay, so let's bring this back around. We, that was a fun discussion of regulation. That's really not the issue. Ultimately, cyber, we kind of got a little bit off from, yeah, we went off the rails, our trajectory. The challenge I see with the Cybermark right now is in the re— just in looking at the requirements. And so I, I, I got down to page 50 of their release document where I think I found the requirements. And what they're saying is they're using NIST IoT product capability criteria, which is fine. NIST writes lots of good stuff. But the challenge I have with this is, I'll give you data protection. Here's what it says. The IoT product protects data store across all IoT product components and transmitted both between IoT product components and outside the product from unauthorized access, disclosure, modification. And then there's a paragraph on cybersecurity utility. The challenge here is we're not talking anything about, there's no, it's like, I, I, I could, I could protect it, I guess, by Caesar Cipher or DES or like, there's no, there's no standard here that says I would expect a set of product criteria to have a minimum encryption expectation. Like if you're not using AES, I think we can agree, if you're not using AES at this point, you're, you're, you don't care about security because there's really nothing else from a, what else would you use from a crypto perspective other than AES? So why not just say, hey, a minimum barrier to a Cybermark product that protects data is the use of the AES encryption algorithm.
32:48Izar Tarandach The proper use of the AES.
32:50Matt Coles Proper use, right. So, so now you're, now you're suggesting, now you're suggesting more than just the guardrail, right? So now you're setting a minimum limit, like how tight, how tight are the guardrails, right? So use strong encryption versus use this encryption.
33:08Chris Romeo I spent a lot of time in my career working in the world of certification. And so I learned a lot about dancing because you can, 'cause I can dance around a requirement and make, I, I can, I can comply with a lot of requirements if you give me room to dance just by the way that I describe what I'm doing and things like that. There's, and that, that's always been one of the challenges in the world of certification is it's not. Ones and zeros. It's this gray area of, let me make an argument for why I actually comply with this. And so in this case, I think you need something. Who in the world of security is going to argue, who's going to say, no, we shouldn't use AES, we should use something else.
33:54Matt Coles You're touching upon something else, and it's the important part of what the Cyber Trust Mark is. Trust. What's the value of the label? If you don't have that minimum. So in other words, if the goal, if the goal is consumers need something they can trust, they look at the label, go, huh, I think I can trust this. Well, why are you trusting it? If the label doesn't include using strong encryption, meaning AES-256 or better, then there's no trust, right? It's just a mark.
34:25Chris Romeo There's no value in the label either.
34:27Izar Tarandach Wait, wait, there's one thing that we haven't touched on. And it's important. Apparently the, the mark always also comes with a check by a trusted lab. Now we can question the validity of that thing and if they're going to positively trust, uh, check, negatively check, whatever, but they are trying to give a very, very specific mark, specific value of trust to carrying that mark.
34:56Matt Coles But that's only if the rules, that's only if the requirements that are the things that they're validating against are meaningful and add value to that trust.
35:05Izar Tarandach But we agreed that they are using NIST and NIST usually puts out good stuff.
35:08Matt Coles Yeah, but NIST has PQC and stuff. I don't think the lab is going to be looking at post-quantum cryptography algorithms.
35:14Izar Tarandach My problem, my problem is different. We, we, we do know the variation in quality and time that it takes to have a good pen test done. So this lab is probably going to be backlogged until wherever, right? And that's going to stifle, to stifle innovation, Chris.
35:32Chris Romeo Yeah.
35:32Izar Tarandach That's not always one of the thing itself.
35:34Chris Romeo So if we look historically, like I said, I worked in certifications, both Common Criteria and FIPS 140 for a number of years. The lab bottleneck could never keep up with the rate of innovation that we were doing. So what did we do? We certified particular versions of software just to get on lists that people use to decide what they could buy. Now, granted, this was at the government level, so we're not, it's, it's a different world that consumer, but consumers aren't gonna be like, oh, well, I can't buy this if it's not on the list. That's the world that we lived in there. But the, there is value in the independent laboratory performing this assessment. The, the challenge is they have to automate it in some way.
36:16Matt Coles It has to be like, uh, well, a lot of this is on, a lot of this is automatable.
36:20Chris Romeo I mean, you could write scripts to check for, no, I'll give you a really, you know, a really geeky example of something that's really cool in FIPS 140. Okay. Probably, it's probably never been recorded on a podcast before, like something cool in FIPS 140. This is really on the, on the deep end of nerdiness, but FIPS 140 has these, I don't know if they do now, they used to, these algorithm tests. And so you basically exposed a testing harness in a test build of, of the operating system for whatever you were having, uh, whatever was the unit under test. And NIST had this, this harness, which would then send various combinations of key and ciphertext, and then you had to return the cleartext. And you, they had all these different derivatives and all these different different collections of data that they would send in and this automated tens of thousands of examples of various fuzzing. Yes. But, but it was, I mean, not, not so much fuzzing because it was known.
37:25Matt Coles Oh, it was a, sorry. It was a good validation suite. It was a certification suite.
37:29Chris Romeo Good values for everything. They didn't give you any junk in it. They gave, they always gave you a good key and ciphertext and said, tell us what the cleartext is or different derivatives of that.
37:38Matt Coles Because it, because it has an, for any given input, it has an expected output.
37:41Chris Romeo Yeah. But, but what I can say as a, as a 28-year security professional, I trust the algorithms that have been FIPS 140, that have been through that FIPS 140 test suite, right? Because you can't fake it. There's no way to fake it. It's crypto, right? Like there's tens of thousands of tests that are being run. You could maybe fake some of them, but you couldn't fake all of them to the point. So if something has that mark, PIPS 140, I trust those algorithms coming out of it because I know the rigor that went into it. Well, so here's the question.
38:14Izar Tarandach Yeah.
38:14Matt Coles So sorry, just to shortcut that a bit. If let's assume for a moment that all the tests that the lab could do could be automated and based on the information that they're providing, it almost certainly could. 99% probably at least. If instead of having a lab do the testing for these types of devices, these are $10, you know, if these are $10, $20, $30, $100 devices, if you had a certification suite that was posted on, you know, on GitHub, for instance, as an open source. So it was, it was open as to what the test validation process was and scripts to run it. And companies could self-certify. Would it hold the same value as going through a third party?
39:02Chris Romeo I would say historically, 'cause there's been self-certification and that's, that's been an issue. It's been a topic of debate in this world for the longest people have been certifying things. And there are self-certifying programs even within the US government for certain types of like communication devices and whatnot. Right. You can run your own test suites and the, and the lab does, plays a lesser role in the independent.
39:29Matt Coles But I think, I think you still, there's still the answer of, can you be trustworthy and non-independent or, well, so alternatively, the, if the, if the test suite is open and companies can run, so, you know, take a dozen product manufacturers from, you know, some, some country somewhere. That want to sell on Amazon and they go through the test suite and they generate data of their validation, could they submit to get the mark? And, you know, I'll just go with, hey, by the way, AI can probably take care of this.
40:07Chris Romeo Yeah, I was thinking the same thing.
40:08Matt Coles So, so, you know, maybe that's a way to, to eliminate the bottleneck of the rate of the guardrail regulation, right? Where the where the regulation can put guardrails so you don't fall off a cliff. Consumers can get some trust out of it. Put a pro— a validation process in place. You don't bottleneck it with a requirement through a, through an independent testing lab. And then make it an innovation-driving regulation, which still gives a, a, a floor to security, which is what we really want out of it.
40:40Chris Romeo The one, the, the one challenge here is How do you trust the results?
40:46Matt Coles That's why you don't— if it's open, if it's open, if it's an open concept of what the validation is and the data can be generated and submitted for the person or the organization that issues the mark, they could run through a validation of that data. And there are ways to, there are ways to secure that process.
41:05Chris Romeo But you can cook that data though. That's the challenge, right? That's why the independent, that's why we, that's why we have independent third-party testing laboratories. Is because they're licensed by the, whatever entity is they're performing the testing for. And if they don't comply with a set of standards, they get their sort of, their valid, their license gets pulled and they can no longer do business in this space that they've built.
41:30Matt Coles Right?
41:30Izar Tarandach Absolutely.
41:30Chris Romeo So they, so they have a, a vested interest in never cooking anything.
41:35Matt Coles Right.
41:36Chris Romeo If anything, they're more strict than they need to be is, is, is the challenge, right?
41:40Izar Tarandach Right.
41:40Chris Romeo Because their fear is that they're gonna lose their license.
41:43Izar Tarandach I'm getting so depressed here.
41:45Chris Romeo You know, that's what we do.
41:46Izar Tarandach No, the, the, the more you guys talk, the more I notice that this whole thing stands on, on, on a, on a, what's it called? A castle of cards. I mean, we are just assuming, yeah, we, we are just assuming trust left and right and up and down and hoping that people are doing the best thing that they can do because they have to keep their livelihoods and whatnot. While at the same time we have the data showing variation of all this, this tests and, and we know how they are limited by budget, limited by time.
42:20Chris Romeo And you want me to depress you even more? We need to depress you even more.
42:23Matt Coles Yeah, go ahead, Chris.
42:26Chris Romeo So there's a whole bottom 10% of any product market where companies just go and, there are, there are places where you can contract with to build you a piece of technology. You just tell them, I want a box, I want it to do this, and I want the cheapest possible COGS or cost of goods sold associated with it.
42:49Izar Tarandach Yeah. Well, that's military logistics, right? You pay the most for the least.
42:54Chris Romeo Go to, go to Amazon and search for, um, HD camera.
42:58Matt Coles Anything, anything really.
43:00Chris Romeo And you will find one for under $10.
43:02Matt Coles Yes.
43:02Chris Romeo And that one that's under $10, I guarantee you was, went through that channel I just described where a company said there is a, people wanna buy these things and they wanna spend the least amount possible so I can build something that's the biggest piece of garbage that barely does what the thing that they want it to do with the cheapest components and the least amount of software functionality in it. And I can sell a billion of 'em. On Amazon.
43:25Izar Tarandach Look, I did the other day, I had to reconfigure my, I have one of those Wi-Fi Bluetooth enabled power things.
43:39Matt Coles Things. Yeah, those things. Yeah.
43:41Izar Tarandach You know, the outlets, the outlet thing.
43:44Matt Coles Oh, a power strip. Yeah.
43:47Izar Tarandach And I use it to not have to get up to my 3D printer to actually press the button and turn it on. So I had to reconfigure it because all of a sudden the system stopped. No, wait, it gets, it gets worse. It gets worse. So I had to try and reconfigure it because my Google Assistant stopped talking to it and wouldn't turn it on when I tell it to turn it on. So that led me to figuring out that the app that was used to configure the thing in the first place. doesn't work anymore. That doesn't support it anymore. Even though there is a QR code on the back of the thing that takes you to Google Market to install the app. Right.
44:28Chris Romeo Okay.
44:28Matt Coles Of course. What, what could go wrong there?
44:30Izar Tarandach What could go wrong? And while I was checking the open source, uh, uh, uh, uh, options to set the, the, the whole thing again, I got to thinking. About the amount of things that I have here in this office that are connected to Wi-Fi. And what if by nature of just configuring them, I am actually sending my Wi-Fi password to a big database in the sky somewhere the east of here. And while of course the first thing that comes to mind is I'm not that important, I'm not that interesting, and nobody wants my particular Wi-Fi password. To do stuff. Once you start having access to those things in volume, then things get very complicated and different at the same time.
45:19Chris Romeo You are the product. Yeah, but that's the, I mean, that is the ultimate challenge in dealing with personal privacy is that people are just like, eh, my stuff's, I'm not that important or whatever. Okay, then why don't you put a drop of blood on a slide, put that in an envelope and send that to that company. But wait, even before— so they have your, your full DNA.
45:43Izar Tarandach Even before we get there, now we, we are jumping into a scheme that, as we said, it's going to put people in Best Buy looking at QR codes and scanning them. And we haven't even touched the, the threat where somebody just prints a number of those labels and goes to Best Buy and just smacks their own QR code on top of the QR code.
46:04Matt Coles That can never, that can never happen.
46:06Izar Tarandach Right? And, uh, like, I mean, this whole thing is so harebrained at the end of the day.
46:14Chris Romeo I think we gotta leave it for now.
46:16Matt Coles Yeah. Let's, uh, let's, we touched upon a part, a lot of deep topics. Yeah.
46:20Chris Romeo Let's, uh, let's let this ruminate for a while. Maybe we'll come, maybe we'll, we'll, we'll pay attention to what happens with this over the next number of months and maybe we can circle back and do an updated conversation a little bit later. But I think for today, I think we have. Uncovered enough rocks, thrown enough stones, yelled at each other enough for one more episode.
46:41Izar Tarandach And getting very, very, very sad about the state of things.
46:44Matt Coles Yeah.
46:44Chris Romeo Sorry, that wasn't our intent, but that's just security. What can I tell you? That's where we live and work.
46:50Izar Tarandach All right.
46:50Chris Romeo Thanks folks for tuning in to another episode of The Security Table, whether you're in Dublin, Ireland or Australia or standing on a treadmill. We hope you have a great day. Haha.
Transcript supplied by AssemblyAI. It may contain transcription errors.
Originally published as The Security Table. Part of the AI Security Table archive.
