Quantum Cyber Security [ML BSide]

Transcription edited by Craig Zorn

[Ran] Hi and welcome to Malicious Life B-Sides, I’m Ran Levi. In our narrative episodes, we usually focus on the past, telling stories of interesting hacks and the like. B-side episodes are where we can usually lift our gaze and focus on the present, and in this case, the future.

[Ran] Quantum computing is a fascinating and revolutionary technology that has been gaining significant ground in the past decade, with researchers from both academia and the commercial sector, such as Google and IBM, announcing major breakthroughs every few weeks. One of the biggest promises of quantum computing is the prospect of revolutionizing encryption. Quantum computers are expected to be able to easily crack even the most advanced classic encryption algorithms.

[Ran] Our guest today, Mike Redding, CTO of Quantropy, a company specializing in quantum encryption, claims that this revolution is even closer than most of us think. He calls it Y2 Quantum and thinks that modern organizations need to start thinking about overhauling their encryption algorithms, much like they were forced to fix their software in preparation for Y2K 20 odd years ago. Before joining Quantropy, Mike was the managing director and co-founder of Accenture Ventures, where he incubated and launched technology innovations for enterprises across multiple industries. He speaks frequently on the impact of emerging technology on large organizations. Mike spoke with Nate Nelson, a senior producer, about quantum encryption and its impact on modern cryptography and the challenges associated with coming up with new encryption algorithms that can stand up to the power of quantum computing. Enjoy the interview.

[Nate] Mike something has been nagging at me as I prepared for this interview. We’re about to have a whole conversation about quantum cybersecurity and yet the fact remains that there are no fully operational quantum computers in the world today. So why should I or anybody else care?

[Mike] I think that’s the classic debate of what do you mean by fully operational? They definitely exist and there’s really kind of two schools of technology in the market today. First is quantum annealers from companies like D-Wave that have been around for the better part of a decade, commercially available, ommercially scalable technology, and more the universal quantum gate style where we see products from the likes of IBM and others, which again, every quarter practically they’re announcing new availability of more quantum bits.

[Nate] Fair enough, but ordinary hackers don’t have access to this technology, cyber criminal organizations don’t, nation states, I’m not sure. I haven’t heard anything about it, but even if they do, they’re not using it yet. So isn’t this a conversation that we should be having five years down the line as opposed to now?

[Mike] A couple different things. A, it is available because it’s on the internet, it’s cloud attached. So you can actually access quantum computers from your laptop or your mobile phone today. So it is available to anybody who wants it and actually there was a study published right before Christmas this year of 600 cybersecurity experts around the world and the consensus is we’ll see the first fully operational quantum attack on cryptography in about two years. So right around the corner and when you think about core systems and core internet technologies, two years is a blink of the eye.

[Nate] I guess, but we’re still speculating on the future.

[Mike] So even though experts are saying it’s two years out, there’s an immediate threat, which is steal now, crack later, because even if it’s encrypted today and the quantum computer doesn’t yet exist, storage is cheap. So if a bad guy grabs your encrypted data, stockpiles it in two years, three years, whatever, they can now crack it open and of course by then they’ll use something like AI to read your data and find the valuable bits that are still valuable. My picture of my dog isn’t going to be valuable in two years, three years, four years, but stealth bomber plans are. And so stealing of data today is vulnerable to those future attacks.

[Nate] All right. And before we get into how quantum cryptography works, Mike, let’s just establish a baseline of what kind of security we have today. Could you tell me in as simple terms as possible how modern cryptography works?

[Mike] Well, modern cryptography is usually a system with really three main parts. The first is trust, right? Establishing trust between any two parties so you know who they are and you know that who you want to do communications with. And to do that, you use asymmetric encryption. Things like RSA have been the industry standard for quite a while. Second, once you’ve established trust, you want to exchange a secret, a key that you’re going to use then to encrypt and decrypt all of your communications. And that key is usually we call it entropy because it’s a random number. And as you suspect, the more random and unguessable or uncalculable, the stronger the key or stronger the password. So trust entropy in the third part is uncertainty, which is the encryption itself, symmetric encryption, which is what you use to encrypt a web session between you and Google, to encrypt your Zoom calls, to encrypt your text messaging on WhatsApp. Whenever you want to encrypt any data that you put in storage, you use symmetric encryption because it’s fast and robust. And so any complete system has good asymmetric, good symmetric and good keys. And if you have all three that are strong, then you’re protected against all attacks.

[Nate] Yeah, quality encryption algorithms today are pretty strong, right? I mean, I always hear about how algorithms would take, I don’t know, like a zillion years for any computer to crack. So why isn’t that enough?

[Mike] Well, so the first vulnerability that comes into play is on asymmetric with that exchanging of the keys and establishing trust. Because unfortunately, I mentioned RSA, that is an algorithm that, you know, in the highest level terms is basically created, you create a public and private key through the multiplication of prime numbers, because that’s, like you said, takes forever for a classic computer to factor so big becomes impossible, practically. The challenge is, in the 90s, Shor’s algorithm was published, which proposed how a sufficiently powerful quantum computer could make easy work, relatively speaking, of prime number factoring, meaning it could open every envelope that was sealed with RSA. And then once you open the envelope, you steal the secret, you steal the key. If I have the key, I can unlock the lock. And so that is what’s expected to fall first sometime in that two-year window horizon.

[Ran] Mike mentioned Shor’s algorithm. And I think that’s a good point to pause the interview and say a few words about the basics of quantum computing and how are quantum computers able to crack modern ciphers so easily. The basic difference between classical and quantum computers is that classical computers use bits, while quantum computers use qubits. What are qubits? Well, a bit can assume one of two distinct values, either a 1 or a 0. A qubit, short for quantum bit, can assume both values at the same time, meaning both a 1 and a 0. That’s a phenomenon called quantum superposition, and if you’re trying to wrap your head around it, don’t. Our minds evolved in the classical world to eat bananas and flee from tigers. We can’t intuitively grasp quantum phenomena more than, say, a monkey can understand C++. And that’s probably more because of C++ and less because of the monkey, but that’s beside the point. Anyway, superposition was around for quite a few decades, but for most of that time, it was a theoretical idea that didn’t have any practical applications. Until 1997, when a mathematician named Peter Shor figured out a theoretical way to use superposition to crack open even the hardest ciphers, ciphers that a classical computer would need many millions of years to break, within literally minutes. No wonder that Shor’s paper ignited a fervent arms race between almost all major powers vying for that holy grail of quantum supremacy.
One obvious question is, how can a quantum computer easily crack a cipher that a classical computer cannot? Well, one physicist, David Deutsch, an early pioneer of quantum computing, has a very interesting theory. Say that I have a pile of a thousand bricks and I wish to build a brick wall. Assuming that laying a single brick takes one minute, it will take me a thousand minutes to build the whole wall. This is much like how a classical computer tries to solve a cipher, with many calculations being done sequentially, one by one, thus taking a long time to complete. But if I had a thousand listeners of malicious life helping me out, each one laying a single brick, I could probably build that brick wall much faster, since many of the bricks could be placed in parallel at the same time, instead of one by one. Peter Shor proved that a quantum computer can do much the same thing, replacing the sequential calculations of a classical computer with much faster parallel computation.
But we know that in our classical universe, we can’t perform all these millions of calculations in parallel. That’s what makes modern ciphers so effective. And if these parallel computations can’t take place in our classical universe, where do they take place then? Well, David Deutsch’s idea is that all those calculations take place in other universes. That is, instead of doing millions of calculations, one by one, in a single universe, quantum computers, quote-unquote, divide those computations between millions of parallel universes, each universe handling only a single calculation in parallel with the rest, much like a thousand people working together, each one laying down a single brick. No, no, no, don’t try to understand how that works. You can’t. Bananas, tigers, monkeys coding in C++, remember? It only means that if Deutsch is right, quantum computers are somehow able to bridge the gaps between these different universes and enlist them to help us efficiently solve a difficult puzzle that is almost impossible to solve in our universe. And if that, ladies and gentlemen, does not blow your minds, I can’t imagine what will. Back to Nate and Mike.

[Nate] OK, then how does quantum cryptography work?

[Mike] That is the grand challenge, which is to come up with an approach to, again, encrypting these and sharing these keys, establishing trust that doesn’t use prime numbers, but in fact does something that is mathematically hard that a quantum computer doesn’t through its superpowers, through its exponential powers. It’s a math problem that doesn’t get sped up and that the exponential powers of the quantum computer can’t crack. And so that’s what NIST has spent since 2015, 2016, has been soliciting the world for mathematicians to submit mathematics that will both do the function of the key encapsulation and the key exchange. But there’s no known weakness from what we can currently contemplate based on the superpowers of the quantum computer. And that’s the competition they’ve been running for a few years. And they’re down to four finalists for key exchange and three finalists for digital signature. And if and when they pick one, that’ll become the new standard with the thought that it will at least last a while until some other mathematician cooks up something new.

[Nate] And that’s useful context, but I’m just really curious about how all of this works under the hood. Mike, is there any way that you could explain the nuts and bolts of quantum cryptography for somebody like me?

[Mike] Well, and that’s where we start to roll out our green eyeshades in our mathematical degrees. Because what it comes down to is today, the envelope, if you would, that you use to encrypt the key, again, this is for the key sharing, is the multiplication or the factorization of large prime numbers. And so now it comes down to how do I have the equivalent public key, private key, where you can never deduce the private key from the public key, and you can never deduce the secret when you see the encrypted packet with the public key. And those attributes require a completely different mathematical basis. One of those main schools of thought that’s currently the subject of some of the NIST finalists is a lattice-based approach, where you basically create a very large, again, matrix or set of mathematical functions, and what number you pick from that mathematical function isn’t really guessable in advance, and there’s no easy way to compute it. And so it’s basically ever more increasing the complexity of the mathematical functions versus just simple multiplication of prime numbers. That’s the giant shift. But that’s also the challenge, because more complex math, the risk is that there’s an unknown weakness, and that’s why they’ve spent five years crunching on it and having the world kick the tires and all the mathematicians that can try to make sure there are no hidden weak points. But also, as it becomes more computationally intensive or more mathematically intensive, it becomes computationally intensive, and that then starts to get itchy when you start to think about IoT, mobile, your ring camera on your front door. How do I protect those things that don’t have the latest and greatest microprocessor running on them? So how do you get to the small end of devices if you have to do a lot of math? And so that’s the constraints. It has to be complicated enough that a super quantum computer can’t bust it, but it also has to be efficient enough that a doorbell can do it. And that’s why it’s taken these five years and a lot of mathematicians, a lot of head scratching to try to come up with something that they believe will survive the next 10 to 20 years.

[Nate] Yeah, it’s really tough for me to imagine how you would even begin to solve a problem like that.

[Mike] Here’s, again, one of the grand design challenges. The internet is 50 years old. IT systems go back 60, 70 years. We have a global install base. And so one of the design challenges is you can’t replace the internet in two years. You can’t replace millions of applications in two years. So quantum proof cryptography or quantum secure cryptography has to look in terms of how it fits in like the classic stuff. It just has to be based on a different style of math so that it can’t be computed with a quantum computer in a way that breaks it. That’s the grand challenge. A few other folks, and this is when you take out an algorithmic approach, which again fits in today’s world. If you can upgrade the hardware, then people are starting to look at quantum key distribution or QKD, which uses photonic entanglement to actually use quantum physical properties of light to basically communicate secretly between two locations and as a result use quantum physics itself. Challenge with that is it’s still a science experiment and, oh, by the way, requires a completely new infrastructure. And again, for governments and big banks and very, very, very, very high secure applications, you may spend millions of dollars on exotic quantum hardware, but for the world economy and the internet as a whole, you got to try to make it work in the rules of today’s road. That’s why this is such a hard engineering problem. There’s plenty of exotic math, but to get it to work in a pattern that fits today’s install base, that’s what makes it tricky.

[Nate] So just to summarize then, we have two independent paradigms. There’s physical security that leverages quantum principles, and then there’s algorithmic security that leverages only mathematics to achieve a state of security that could beat even quantum computers.

[Mike] Right on. Right on. So one’s an algorithmic approach, which is your computer approach, the right code that fits in today’s operating model so you can do an in-the-field upgrade. The other requires a whole new infrastructure, and that’s still a long way away from being commercially viable.

[Nate] So are you saying that we are closer to the former than the latter?

[Mike] Oh, definitely. With the algorithmic approach, NIST is down to seven finalists, right? They’re close. They were actually due to announce the winner or the finalist candidate in December. It’s February, and they’re a little slow, but they’re trying to be right because when they set the standard, the whole internet is going to shift in that direction. So they’ve got to be right. That’s why it’s taken them five plus years to get this far, but the hardware one is exotic and exciting because it starts to talk about you’ll see the buzzword quantum communications and a lot of other sexy stuff. But even the NSA has said, for now, it’s a dead end. And so the hardware will come someday, but not soon enough.

[Nate] The reason being that what? It’s just too difficult to work on a quantum physical level.

[Mike] We just haven’t invented it yet. I mean, look at my best, you can talk about analogy. The best analogy there is nuclear fusion, right? We’ve known about nuclear fusion since the fifties, right? Yet to this day, 70 years later, and we all know it’s clean and green and super powerful. It’s what powers the sun. So, hey, that’s pretty cool. 70 years of physics and engineering later, we’re still working on it. Well, physics is hard. And so with the quantum physical systems, physics is hard. And so we see the potential, but an actual practical, scalable, affordable, secure years away.

[Nate] I totally follow what you’re saying with regards to why physical QKD is so difficult to achieve. And yet, really, as I think about it, the algorithmic approach is even harder for me to imagine, you know, that anything that could fit in my doorbell could stack up against the power of one of those massive quantum computers.

[Mike] Well, one is the weapon and one is the shield, right? And the little ring doorbell just has to have enough of a mathematical shield that no matter how much horsepower, right, you know, no matter how much brute force quantum mathematics and quantum compute, you know, firepower you bring to bear on it, your math is elegant and rock solid and cannot be computed. Then no matter how much, you know, how many atoms in the universe you line up against it, the math holds, right? That’s the purity. That’s the beauty of mathematics, right, is that if you find the right algorithm, it may be unsolvable, right? And in fact, there are certain classes of algorithms that are used in this that are being considered for this, that, you know, there’s 200 year old theorems in mathematics that say they are unsolvable, no matter, it’s just mathematically unsolvable. So that’s the class of problem people wants to use. So that’s why the encryption is small, even if the attack is big.

[Nate] All right. So who needs this right now? Are we talking about regular folks like me and you or are we talking about like the military?

[Mike] We always start with, of course, the military and banks because that’s money and state secrets. But arguably what I share with a friend or a colleague, it’s personal and it’s private. And I don’t necessarily want anybody else to know it. So I may have, you know, less economic incentive than say a government that’s trying to protect stealth bomber plans, but it’s not, you know, people, you know, I don’t want my password stolen. I don’t want my money stolen. I don’t want everyone to know my business. And you can even see, I saw it again on this weekend watching sports commercial for one of the messaging platforms that basically went out and said end to end security. No one can ever read your messages. And this is definitely one of the biggest, you know, social media messaging platforms has billions of users and they’re advertising. Nobody can read your messages. So clearly they wouldn’t be putting it on during a sporting game and spending a lot of money on ad if they didn’t think that was a buyer value. It’s a mass market theme, let alone, of course, government banks, you name it, or things like even like the modern connected car. You know, you don’t want anybody turning your car off while you’re driving down the highway, right? Whether it because they’re malicious or accidentally, right?

[Nate] Okay. Now, I know that some governments and large companies are working on quantum computing, but what about quantum security specifically? Are there any countries out there or major companies that are taking this issue really seriously that are worrying over it and investing in it right now?

[Mike] Absolutely. I mean, everybody’s taking it very seriously. You know, all the major governments around the world, you know, from a Western perspective, at least leads the charge. But a lot of the standards when it comes to cryptography are intertwined, for example, between the EU and the US, right? They kind of refer to each other because, again, we are trading partners and allies and a bunch of other stuff. And not surprisingly, different ecosystem, China has some of their own standards and some of their own intent because they want to protect their economy in their own right. So yeah, it’s a multipolar world. So we definitely see a couple of different, you know, major locus of activity. Western, Eastern, and you name it. So yeah, they’re the government. But then every major tech player, again, starting from the kind of a US centric view, but the IBMs, Microsoft, Google’s, all of them are putting a lot of horsepower, again, not just into the creation of the compute, but definitely into the creation of the security. Because again, all of those companies, the tech giants make their money from online, whether they’re public clouds or software platforms or you name it, they make their money. And if the internet crashes down and the digital economy crashes, so does their market cap. So they’re definitely elbows deep throwing their very significant resources into the community to make sure that what comes out works for everybody.

[Nate] So are there any solutions, any technologies that have trickled down that are available to us right now and attainable today that can help organizations or folks that want it get ahead of the game?

[Mike] Well, I would say being a student of the NIST process, knowing that everything that they’re doing is on the internet and it’s open source. So arguably, if you want to get started today, go download the software, get it in house,
start kicking the tires, start seeing how it behaves pretty much like you’d expect your legacy stuff, your legacy encryption to work. So again, that’s part of it is that it’s not meant to be a whole new way of thinking. It’s meant to more be new code that replaces old code, right? But it should work in kind of the same pattern, the same application pattern, the same calling pattern that what you’re used to. So it’s not meant to be some exotic, you got to learn a whole new skill set. It’s more like I got to just use a different set of libraries.

[Nate] Then what you’re saying is quantum cryptography is something that is attainable for any organization. You could implement it right now.

[Mike] But the grand challenge is that for decades, crypto has been below kind of the line for so many companies because it’s settled. We know which encryption we’re using and we may not even know it because it’s embedded in the products we use. So we like to call it Y2 quantum. It’s kind of like Y2K, 20 odd years ago, you knew that you had to change the dates in your computer code and you had to go through all your computer code to find where you’d done it wrong and taking a shortcut and had to put the longhand version in. But you knew when that was going to be due, right? You had a deadline. With Y2 quantum, it’s mushy, right? It’s two years plus minus. But you got to go through all your code and say, where do I use crypto and where and therefore where do I got to do a patch, right? You saw it a couple of weeks ago with log4j where everyone had to go back through all their systems, all their systems. Hey, where did I use that? I didn’t even think about it because everybody uses it. It’s common practice. You don’t even have to think about it. Whoops, got to go back and think about it. So the bigger challenge is just nailing down the hundreds if not thousands of places a given company or government agency uses cryptography so they can make sure they upgrade them all. That takes probably more of the time to be aware of cryptography and where it’s deployed versus the actual upgrade itself.

[Nate] Michael, is there anything left to say before we part ways?

[Mike] Well, I think it’s important to be an advocate, right? To be aware of the trends in quantum computing, which again, are going to have so many benefits for mankind that it’s super exciting, but understand that it’s got a security threat through as well. And so just starting down the journey, starting to be aware, starting to do that inventory, starting to kick the tires and get hands on with it. Now, when it’s not dire, when the bell rings and we have the first case of a quantum attack breaking somebody’s key, the rush is on. And don’t be the last person in line to buy your inoculation. Be the first person, know what you’re doing, build your immunity, and that way you don’t got to worry about it.