Hart Montgomery 0:00 More than 90% of closed source software had open source software dependencies. We recently saw a study that said not only was that the case, but that 90% of closed source software in this particular survey that that they did, which was thought to be representative, had software dependencies with potential security issues, which is even more frightening.
Johannes Lintzen 0:25 Welcome to Shielded, the last line of cyber defense by PQShield, the podcast where we are moving the conversation around post quantum cryptography from the why to the how. I'm Joe Linson, your host, and I'm here to guide you through the challenges, innovations, and solutions that will help your organization stay secure in the quantum era. Let's future proof your defenses together.
Unknown 0:50 Today, it is my pleasure to speak with Hart Montgomery, technical director at the Linux Foundation. We will be talking about three of my favorite things, burgers, beers, and crypto agility. Listen in and learn how to climb up the steep hill of quantum migration. Today, we're joined by Hart Montgomery, technical director at Linux Foundation. From his recent findings in critical financial infrastructure to the path toward crypto agility, Hart brings insights from the front lines of post quantum security. He'll share what organizations need to know about risk assessment, implementation challenges, and strategic approaches to this complex transition. Welcome, Hart. How are you doing today?
Hart Montgomery 1:35 I'm doing great. Thank you very much for having me today.
Unknown 1:37 Absolutely. We're excited to have you. So Hart, when we first spoke, I understand you live in Bernal Heights, close to what many consider one of the best breweries on the West Coast. So maybe you can share a little bit with some of your favorite beer.
Hart Montgomery 1:51 Oh, that's right. I've always been a big IPA fan, and there's a brewery right nearby called Bear Bottle that has a lot of the best IPAs, I think, in the Bay Area. So my wife probably thought it was too big of a factor for me in moving while we moved to this proximity to this brewery, but I always enjoy it, and I always enjoy a very strong and happy IPA. Yeah. And you're
Unknown 2:15 in walking distance, I suppose, so that makes it even easier.
Hart Montgomery 2:19 It's about two blocks. It's about a 40 degree hill on those blocks, but that's Bernal Heights for you. to this brewery, but I always enjoy it, and I always enjoy a very strong and happy IPA. Yeah. And you're
Unknown 2:15 in walking distance, I suppose, so that makes it even easier.
Hart Montgomery 2:19 It's about two blocks. It's about a 40 degree hill on those blocks, but that's Bernal Heights for you.
Unknown 2:24 Right. It has it in the name, I suppose. So I guess brewing beer is a lot about consistency. I mean, that's what the good breweries separates from the rest. They just crank out very consistent product. And I think in cryptography and cybersecurity, consistency is is also key. I found it interesting that you have a PhD in cryptography and you focused a lot on scalable, secure blockchain architectures. And before joining Linux Foundation, if I'm not mistaken, you were a blockchain researcher at Fujitsu. So I'd be curious what sort of motivated you, you know, drew you to the field and if your academic background and I believe your PhD even is in lattice cryptography, so quite fitting with the post quantum cryptography discussions that we're having. So I'd be interested how your academic background also influenced your approach to this whole discussion around quantum readiness.
Hart Montgomery 3:15 Absolutely. I mean, I think having an academic background in this space is extremely useful. It's not necessarily for implementing or using cryptography, but sort of understanding all the core mechanics, how things work in detail is extremely useful. And I've always been someone who likes to see cryptography get applied. Math is beautiful and everything, but there's a different perspective entirely when you want to work on things that actually make it to the real world and actually get real world use. That's why a lot of cryptographers have worked on blockchain in recent times because it is a way to get cryptography used in practice oftentimes very, very quickly. And I really credit a lot of the big blockchain communities like the Ethereum community, for instance, for really being progressive with adopting new cryptography and cryptographic techniques. So naturally, a lot of cryptographers have been drawn to that space, particularly the ones that care about seeing cryptography work in the real world. Now we're seeing that in the post quantum space too, which is really exciting.
Unknown 4:25 Well, I mean, speaking of real world examples and the things that you shared with me sort of preparing for this podcast, you did mention that you recently were involved in security audit at a large unnamed financial organization. And I think you used the word scary to describe the results. So I'm sure our listeners would be curious to hear what the most concerning findings within that report. d with me sort of preparing for this podcast, you did mention that you recently were involved in security audit at a large unnamed financial organization. And I think you used the word scary to describe the results. So I'm sure our listeners would be curious to hear what the most concerning findings within that report.
Hart Montgomery 4:48 Oh, man. And I've seen another report recently. There's a big open source study. So we used to always cite at the Linux Foundation that more than 90% of closed source software had open source software dependencies. But we recently saw a study that said not only was that the case, but that 90% of closed source software in this particular survey that that they did, which was thought to be representative, had software dependencies with potential security issues, which is even more frightening. But I've seen a number of of big orgs. I wouldn't say I'm involved in the security audit directly, but I've talked to people about results. And it is frightening. You see things like MD five certificates, just really, really old cryptographic code. It's it's sort of like, why does stuff exist? Somebody yesterday was telling me about how they found a single DES in somebody's code base, like, ten years ago, which is like, okay. That's pretty bad. So a lot of this is just organizations not necessarily having the cryptographic expertise or the security expertise to properly put in place policies and really enforce updates. If you have an org where a dev can just kinda go pick out their own cryptographic algorithm and put it in and use it, then you're gonna have problems.
Unknown 6:13 So it it sounds like the migration towards quantum safe cryptography really isn't from a baseline of what is considered best practice today. It's, in some cases, migrating from something that was considered best practice ten, fifteen years ago. And I think financial organizations are probably the best example because they even using still triple DES in some of the arguably, I mean, payment transactions are short lived, so I mean, there's less of a challenge there. But, I mean, triple DES has been considered obsolete for fifteen years now, something like ago. And I think financial organizations are probably the best example because they even using still triple DES in some of the arguably, I mean, payment transactions are short lived, so I mean, there's less of a challenge there. But, I mean, triple DES has been considered obsolete for fifteen years now, something like
Hart Montgomery 6:46 that? Or in some cases, not using best practices altogether or in some cases, just not being consistent. I mean, you really hit on a good point here, and, I mean, this gets into cryptographic agility, which everybody talks about these days. If you have a crypto microservice or library and everybody in your company uses it, then it's much much easier to potentially do a migration. Very, I'd say, crypto forward organizations have this kind of thing, like a lot of the big tech companies like Google, AWS, where migration is really just somebody on the security team or a committee of people on the security team, really, changing a parameter to an algorithm. Obviously, there's a lot of consideration that has to go into that because there could be breaks and unintended consequences. Maybe some protocol is reliant on a signature plus some other stuff fitting in a packet. And then if that doesn't happen, then something unexpected happens. But that kind of thing is, like, much more straightforward than people that have sort of hard coded ECDSA into their code base everywhere or specific hash functions with very specific input sizes. So that kind of thing is much, much harder to migrate.
Unknown 8:02 And I think you'll use the term before. Right? Policy driven approach to this, of course, would be ideal. And then if you had something centralized that, like, a crypto as a service solution almost, that then doesn't allow the developer necessarily, like you said, to choose their own hashing algorithms or signing algorithms. They just have to follow policy, and and that gets put into place by role based controls. And then I suppose how would we bridge over to quantum readiness within such a centralized crypto as a service solution? How would a migration What could a migration look like if there was such a thing as a centralized control? s. They just have to follow policy, and and that gets put into place by role based controls. And then I suppose how would we bridge over to quantum readiness within such a centralized crypto as a service solution? How would a migration What could a migration look like if there was such a thing as a centralized control?
Hart Montgomery 8:39 Well, it's much easier if you have some kind of crypto microservice or general architecture. So if you look at sort of recent publications, I would say not formal publications, but blog posts or other things by Google and Amazon. I'll use them as example. IBM also has some stuff on this. They really segment cryptography applications into a number of different types. Basically, things where PQC migration is very inexpensive in the grand scheme of things or things where it's more expensive. So for instance, for signatures, the biggest drawback with the post quantum migration for signatures is the signature sizes are generally well, they're always larger. Right? Pretty much. And the keys can be larger. So in some cases, this doesn't matter. And it's a similar thing for public key encryption or key exchange. I'm sure there's most people who work with this a little bit know this. But for instance, if you're running TLS or SSL or something, the initial key exchange at the very beginning is a very, very small part of most sort of traffic flow. So if you have a little bit of extra overhead, if you have a few 10 kilobytes even, say, of overhead and your whole exchange is in the megabytes, well, it kind of just doesn't matter. So we've seen a lot of adoption for that for things like messaging where it's also relatively inexpensive, where we've seen signal and Apple already sort of push through these migrations. So for things like that, it's easy. And for some other things, it's, like, quite expensive, like for certificates. If you're dramatically increasing signature sizes, then you're just dramatically increasing certificate size. So stuff like that is a little bit more complicated. But if you have these, like, nice microservices, you can just switch with the single parameter, you can just switch these things one by one, make sure everything's okay, and do sort of a very easy migration. Well, somebody yesterday, you know, whose the comment was all migrations are hard. Some migrations are easier than others. So easy for migration. microservices, you can just switch with the single parameter, you can just switch these things one by one, make sure everything's okay, and do sort of a very easy migration. Well, somebody yesterday, you know, whose the comment was all migrations are hard. Some migrations are easier than others. So easy for migration.
Unknown 10:41 Well, I think you're laying out the less resource intense things like TLS handshakes. I think that's also the reason why we're seeing the first migration steps happening within exactly those use cases. Because as much as I don't like the term, but that seems to be the low hanging fruit for PQC migration to go, okay, let's exercise that through and see what results we can get. And I think it's interesting that you mentioned AWS and Google and IBM as sort of examples of, well, they've solved some of these issues for you. So to me, that is the digital supply chain for the large organizations that they're using those services. And I think it's always interesting to look at, well, what are the capabilities that like, if I'm a large organization and I'm tasked with that migration, like, what are the capabilities? What are the questions I should be asking my vendors, my providers as I evaluate those solutions and see if they might be a fit for my specific use case?
Hart Montgomery 11:35 That's a great question. And I think those companies are sort of even more unique because they have bunches of very good cryptography PhDs, which is not typical even in a technology space or especially like a financial firm space. And mostly, like, particularly Google and Amazon, they're focused on their own services, and I think they do a very good job of that. So they've made all these sort of migration priorities, and it's like, well, we need to secure this. And it's really interesting to sort of see their decision making because they put a lot of thought into this stuff.
Unknown 12:09 So listening to your explanations, I think we can say, well, cryptographic agility is key. Centralized crypto as a service type offerings can help with rotating keys and making the transition much more seamless and less painful. In general, I think the whole migration to me feels like we have to walk up a steep hill in Bernal Heights to eventually get to that state that we want to be in with protecting our environments from the quantum threat. I think it's probably historically speaking, and also for this migration, it's fair to say that regulation has been, uh, you know, much of a driver for this. And then I think PQC migration is no different. Do you see compliance as a catalyst for action, or are the global standards moving quickly enough to support those organizations that are making the early moves? I'd be curious to hear your take on that aspect of the whole migration. , you know, much of a driver for this. And then I think PQC migration is no different. Do you see compliance as a catalyst for action, or are the global standards moving quickly enough to support those organizations that are making the early moves? I'd be curious to hear your take on that aspect of the whole migration.
Hart Montgomery 13:05 Yeah. Absolutely. I mean, regulations are driving a lot of this. There are some very sort of security forward and security conscious companies. Certainly again, the big tech companies that employ legions of security and cryptography PhDs tend to be very aware of things and very aware of the risks. That's not the norm in the broader industry. And we do see a lot of people who who sort of you know, it's just like, well, I just wanna make sure I'm okay with the government regulations. The US government has been very aggressive about requiring post quantum protocols in the future. So, yes, that is absolutely driving adoption, and anyone that wants to contract with the government or work on government stuff is definitely paying attention to this. So I think that's a really huge driver.
Unknown 13:52 I agree. And and, I mean, besides the US federal government, I think just today or yesterday, UK released their version of the guidelines. They too have 2035, kind of that ten year window that they push forward. I believe Australia is even a little bit more aggressive with those timelines. Now the other thing that I was reminded of as we were talking about open source being part of a closed source software in 90% of the cases, of course, one way to approach it is the software bill of materials to sort of have that list of dependencies and be clear about it. And closely related to that, we have the CBOM, a cryptography bill of materials, that addresses which algorithms are being used and such. And I believe you mentioned that maybe IBM did a large open source project together with the Linux Foundation around CBOM. So that certainly helps to address the quantum readiness as well. Maybe you can speak to that a little bit. cryptography bill of materials, that addresses which algorithms are being used and such. And I believe you mentioned that maybe IBM did a large open source project together with the Linux Foundation around CBOM. So that certainly helps to address the quantum readiness as well. Maybe you can speak to that a little bit.
Hart Montgomery 14:48 Yeah. And that's really about helping organizations find and locate where they potentially have non post quantum cryptography. And the bigger picture, SBOMs are, you know, completely essential at this point. It's hard to trust a large software project that hasn't gone through the work of putting together an SBOM, particularly if there are a lot of dependencies. We sort of refer to it as the open source hamburger where modern software development is basically done in the following way. You pick a framework. That's one of your buns. It's usually an open source framework. You do a little bit of custom code, which is sort of the meat, and then the other bun is all the open source libraries. So you pull in and used to solve problems. And it turns out the custom code that most people write for these modern software projects is is, like, quite small. It's, like, maybe 20% of the code in an overall project. But, of course, when you do things like that, you have to have SBOMs particularly well so people know what they're getting. So if there's a security vulnerability report on something, you know, and you can easily update the required component. We see this all the time where I'll go through and just take a deep dive on a big project, and it's like, well, you're pulling in this code from this GitHub repo. There's one guy maintaining it. He hasn't touched the code in four years. Like, are you sure you want to do this? So it's these kinds of questions. C bombs are also super important for this post quantum transition. And I think in general, you want to make sure that people are using up to date cryptography. You wanna make sure that that people are consistent on having this information available to you without having to do your own sort of deep dive into the code is extremely useful and important.
Unknown 16:34 Right? Well, and that kinda goes back to your earlier comment on the security reviews that you've seen within the financial organizations. Right? And I'm pretty sure that was eye opening for many, probably even up to the CISO level. They went like, wait. We're still using MD five? I had no idea. Great. But we're finding all these issues now, and that's certainly gonna help to do the risk assessment and to say, well, where's our most critical data? Where do we start with this? I am. that was eye opening for many, probably even up to the CISO level. They went like, wait. We're still using MD five? I had no idea. Great. But we're finding all these issues now, and that's certainly gonna help to do the risk assessment and to say, well, where's our most critical data? Where do we start with this? I am.
Hart Montgomery 17:00 Yeah. Absolutely. And it goes back to the microservices and crypto agility point. Ideally, you'd like your developers to understand cryptography in sort of a black box way. So what I mean by that is they understand what a digital signature is and what it does. It gives you authenticity. They understand what key exchange or public key encryption does. But it's not reasonable to expect everyday developers to understand the math behind all these systems or to be able to decide which one to choose or anything like that. It's like, well, should I use ECDSA or Schnorr? Well, if you're a developer, they're probably indistinguishable to you. So it's not fair to ask devs to make these decisions. But what you can do is you can say, okay. Call this function for your signatures. We'll always make sure we have high security efficient implementation in there, and you can just do that in general. And there will be some exceptions, but to get an exception, you kind of need to know what you're doing and have a good justification for it.
Unknown 18:01 Yeah. And ideally, have two or more people take a look over it and ideally with some deeper knowledge of like you said, a developer might be really good at developing code, but they might not necessarily be really versed in the latest and greatest cryptographic algorithm and how to use it. And also which implementation, quite frankly, which implementation to use because typically, it's not necessarily the algorithm itself that breaks. It's a poor implementation that will be under attack and will eventually allow the bad guys to come in. Absolutely. Yeah. So a lot of organizations that we're talking to, they have a challenge of identifying that initial use case and sort of getting that c level buy in. Is there any recommendations you can give to say, what's some compelling business case to bring to the boardroom and get accepted to get started with, get some quick wins? So a lot of organizations that we're talking to, they have a challenge of identifying that initial use case and sort of getting that c level buy in. Is there any recommendations you can give to say, what's some compelling business case to bring to the boardroom and get accepted to get started with, get some quick wins?
Hart Montgomery 18:49 Absolutely. So the example I like to bring up is the classic harvest now, decrypt later problem. So what this means is obviously, you know this, but, basically, however you're using data, it's probably being transmitted across the Internet today. And people are recording this data, whether it's nation state adversaries or competitors or whoever. People might write your data down. Now the problem is if I write your data down today and a quantum computer comes out five years from now and you use some kind of elliptic curve based key exchange to set up your secure connection, well, I can go back and decrypt all that data now. And if you needed that data to be secure for longer than five years, you have a problem. So this becomes a game of sort of how long does my data need to be secure versus when do I think a quantum computer will be ready? The latter is a difficult problem. Like, people ask me this all the time, but I am not a physicist, so I have no idea. I'm just the same. You gotta ask the physicists, and that's a speculation game. Every year, there are a couple of organizations that sort of poll all the physicists and say, like, in every sort of interval of five years, what's the percent chance you think a quantum computer will be powerful enough to crack RSA twenty four eight or something like that? I think those are really interesting and useful. Again, we don't know for sure, but that's sort of the best we have. But the other side is much easier to sort of guess or estimate. Like, organizations generally should know how long their data needs to be secure, and that's unique to many organizations. But this is the general calculation. The wild card is also trusted hardware. If you need cryptographic supported hardware, then this just throws a whole another sort of epic on the timeline because you need to decide a protocol, fix the hardware, manufacture the hardware, have it out, all sort of in time before you need to add that to the length that your data needs to be secure. So that's probably the most worrisome for me is reliance on trusted hardware, and can we have that ready in time? the timeline because you need to decide a protocol, fix the hardware, manufacture the hardware, have it out, all sort of in time before you need to add that to the length that your data needs to be secure. So that's probably the most worrisome for me is reliance on trusted hardware, and can we have that ready in time?
Unknown 20:57 Well and then that really ties back into the critical infrastructure. Right? I mean, we use financial as an example for critical infrastructure, but, of course, there's things like power supply, water supply. There's, like, millions of devices deployed out in the field. That is a huge challenge to upgrade all of them, and no one's gonna go in and rip and replace because, I mean, that is just cost prohibitive. But coming up with smart ways of migrating those long lived devices out in the field. Even more important to get started with the planning now if you look at those timelines of five to ten years. And then it becomes about a game of balancing, like, your immediate security needs and sort of your long term planning. Do you have an example for sort of that balancing act maybe?
Hart Montgomery 21:39 That's a tough one. And I am very, very worried about sort of the IoT and, as you put it, the small cryptographic hardware pieces. So, like, your power meter, you might have a smart meter. That's gotta get updated and replaced. All the little things like that are exceptionally problematic. The power company is definitely not going to have cryptographers working for it. So this is probably what I'm most worried about. Are these non tech associated or tech adjacent hardware devices that are updated very, very infrequently. And some of them are extremely critical, so I'm not sure there is a good solution for this. Some of them, there's no choice but to rip and replace, so it's problematic.
Unknown 22:23 So tying the conversation back into sort of the Linux Foundation activities, and I know and appreciate you joining us in such a busy week with the member summit taking place. You did mention some interesting takeaways from the keynote. Maybe you wanna share that as well as it I think it ties nicely into authentication and the discussion about open source projects. e Linux Foundation activities, and I know and appreciate you joining us in such a busy week with the member summit taking place. You did mention some interesting takeaways from the keynote. Maybe you wanna share that as well as it I think it ties nicely into authentication and the discussion about open source projects.
Hart Montgomery 22:44 Absolutely. So one of the big things that has been on the Linux Foundation's mind recently are software supply chain attacks, obviously. We talked about SBOMs. In particular, this, like, Etsy utils attack that happened not too long ago where a person operating or people operating under the identity, Zhiyatan, found a sort of critical project in the open source supply chain that was also under maintained, volunteered to help maintain it, and then installed malware essentially to this project to try to compromise a huge amount of our digital infrastructure. So this is extremely worrisome, and it's like, how do we adopt this kind of thing? What can we do? And one of the proposals was, well, this wouldn't have happened if we only sort of allowed trustworthy people or people we could identify as trustworthy with some chance to contribute or maintain these projects, especially maintain the projects. So the idea was that, well, if we think about what the kernel does, the Linux kernel requires people if you wanna be a maintainer, you have to go meet other colonel maintainers and show them your, like, government ID and documentation. This is an extremely tedious process for them. Sometimes, like, prospective maintainers will have to go drive hours to go meet people at different cities, so not the greatest for them. And the idea is that with advancements in digital identity and digital credentials, we can sort of do all this digitally and keep it privacy preserving using zero knowledge proofs. And, of course, this won't totally stop the problem. A super dedicated person could obviously spend years and years and years building up a reputation before doing something. And we jokingly accused Greg Crow Hartman, who's, like, the number two guy on the Linux kernel of being a you say, what if, Greg, what if you're a long term spy and you've built up this position in the kernel to launch an attack? But we can hopefully mitigate a lot of these problems where just sort of low reputation, high risk people are given access to some of the most important code running a huge amount of our digital systems today.
Unknown 25:02 Well, that goes back to the beginning when we were talking about consistency. If you're the protector, you have to get it right each and every time. If you're the bad guy, you only have to get it right once. That's all it takes. Well, Hart, look, this has been a fascinating conversation. Really appreciate you taking the time out of your busy schedule. Before we close-up, do you have any sort of best practices for organizations just getting started on this? Like, what are the top three things to get started with today to sort of be ready? akes. Well, Hart, look, this has been a fascinating conversation. Really appreciate you taking the time out of your busy schedule. Before we close-up, do you have any sort of best practices for organizations just getting started on this? Like, what are the top three things to get started with today to sort of be ready?
Hart Montgomery 25:31 That's a great question. So I would say, number one, even if you don't have a plan, have a plan for a plan. You can call it concepts of a plan if you like. That's
Unknown 25:40 very meta.
Hart Montgomery 25:43 So just think about, particularly for harvest now, decrypt later, how long your data needs to be secure and when you think quantum computers might be viable for breaking cryptography. The second one I would say is start reengineering your crypto code now to be crypto agile. People always ask, what should I do now? And I always tell them, like, building crypto agility tier systems. It's best security practice anyway, so you should be doing it. And you don't even have to say that, you know, oh, look. This is for a a post quantum migration. You can just say, this is so we're secured today, and we're following best practice today. And third, I would say is think about any sort of unforeseen effects of using cryptography in sort of a non black box way. I alluded to the example earlier of some protocols were designed around, like, a signature fitting into a single packet. And, obviously, having a much larger signature would break that assumption or tread. So just think about also those things where you're relying on cryptography having some particular property and make sure that that doesn't change and you don't need to rearchitect your system around that.
Unknown 26:55 Fantastic. Thank you. That's very valuable. And then lastly, before parting ways, do you wanna share with our listeners where they can they can follow your latest work, like latest and greatest that's happening for you at the Linux Foundation?
Hart Montgomery 27:09 Sure. We have a bunch of projects that work on and touch cryptography at the Linux Foundation. The Post Quantum Cryptography Alliance is the one directly related to post quantum cryptography where we have things like the OQS, the Open Quantum Safe Project, which many people are familiar with. We also have Linux Foundation decentralized trust, which has a number of cryptographic protocols in it, a lot of stuff around MPC threshold crypto. The Open Wallet Foundation is another project that works with things like anonymous credentials and stuff like that. So check those out. We have a lot of code, and we're the Linux Foundation. We're open for everyone to contribute.
Unknown 27:48 Fantastic. Thank you so much, Hart, for sharing your insights, and this has been a great conversation.
Hart Montgomery 27:53 Thank you very much. Thanks. als and stuff like that. So check those out. We have a lot of code, and we're the Linux Foundation. We're open for everyone to contribute.
Unknown 27:48 Fantastic. Thank you so much, Hart, for sharing your insights, and this has been a great conversation.
Hart Montgomery 27:53 Thank you very much. Thanks.
Johannes Lintzen 27:56 A quantum era is already here. With the right tools, the right knowledge, and the right mindset, together, we can face it head on. To find out more about PQ Shield and how our cutting edge solutions can help you secure your data against the quantum threat. Visit pqshield.com. Make sure to search for Shielded, the last line of cyber defense in Apple podcasts, Spotify, or anywhere else you get your podcasts. Don't forget to click subscribe so you never miss an episode. On behalf of the team here at PQShield, thanks for listening.