Andrey Maksimov: Today with us is Robert Bedington, CTO and co-founder of SpeQtral, a company based in Singapore, formerly known as S15 Space Systems. We continue our #knowyouroptions conversation with entrepreneurs working on amazing products. And today we’re at the Small Satellite Conference in Logan, Utah.
AM: Robert, tell us about yourself, your background, and then let’s discuss what you guys are doing.
Robert Bedington: Sure. So, my background. Originally, I was a physicist. I worked in space plasma instrumentation, got my PhD making miniaturized instruments. From there I moved to JAXA and was working on miniaturized plasma instruments there. And then, nearly five years ago now I moved to Singapore and started working with a group at the Centre for Quantum Technologies at the National University of Singapore (NUS), where they are making miniaturized entangled photon sources for CubeSats.
AM: Can you explain what it is?
RB: It is not straightforward. An entangled photon pair source makes pairs of photons, or particles of light, that are correlated in ways that classical physics cannot explain. So, these pairs of photons have a ‘spooky connection between them’ – this is how Einstein described it.
AM: That’s how you named your first satellite?
RB: Right. Our first satellite is SpooQy 1 and our mission patch shows Einstein scared of the entangled ghosts here.
AM: I like it.
RB: So, these entangled photons, this property has various uses in technology. And the most immediate use is in the generation of encryption keys, very secure encryption keys that can be shared between two remote parties with tests verifying that there’s no eavesdropper listening in…
AM: So it’s not pure R&D, it’s not pure science, it actually has a commercial application to it, right?
RB: It’s both. So there’s a big kind of scientific aspect to this, but there is also an immediate technological commercial use case. And that’s what we’re exploiting at SpeQtral.
AM: So basically you use the quantum effect or quantum entanglement and try to apply it to the modern cryptography. Is that what you do?
R: Yeah. It’s actually not so modern. The quantum key distribution, this technology for generating encryption keys was developed in the 1980s and then the entangled version in the early 1990s. It’s just only recently that it’s become practical to implement this on a commercial scale. And also only recently there’s been a real push to adopt this technology. And that push also comes from quantum physics, and this is quantum computers.
Quantum computers have received a lot of funding, a lot of research recently. Some people charting the performance of quantum computers say this is moving faster than Moore’s Law. And the reason this is driving QKD sales is that quantum computers we know will be able to crack conventional encryption methods once they get powerful enough.
So, the algorithms required to break public-key encryption schemes, particularly RSA, which is the most commonly used encryption method, the methods to break this have already been worked out. The quantum algorithms that you’d need to run on your quantum computer were worked out in the ’90s. And so we are now just waiting for the machines to get powerful enough to run these applications and then there’s a real security problem.
AM: What is QKD?
RB: It’s quantum key distribution. Quantum key distribution is a method of making encryption keys using photons rather than using mathematical tricks. So, public-private key encryption is incredibly convenient, but it’s based on mathematical algorithms. It’s based on the assumption that the mathematical problem you use to encrypt your message is too difficult for a conventional computer to unpick. And that’s basically true for modern computers, but we know it won’t be true for quantum computers in the future.
AM: So, basically you’re developing a solution for the future world where the quantum computing is so spread out that people can actually use that to crack the codes of the traditional encryption keys, right?
RB: Yeah, that’s one use case that has certainly been driving investment and improvements in QKD. But there are other reasons why you want to use QKD. I mean, it’s provably the most secure method for producing encryption keys, so for the people with the bigger secrets in the deepest pockets, this will always be the most attractive way for them to make encryption keys regardless of the progress of the computers.
AM: Why satellites? Why do you put this technology on satellites?
RB: Why satellites? Because you’re talking here about sharing individual photons and the key is encrypted in individual photons, streams of individual photons. We have to identify each individual photon one by one to make a key from this. And we can’t amplify these signals. So, the key is encoded in quantum states of these individual photons and quantum physics just does not allow you to make copies of this, because it breaks the whole principle. This means that you’re very sensitive to losses. You can’t afford to lose these photons you have to transmit them. So this means that if you’re sending them over optical fibers, which is a normal way to send them if you’re communicating optically, there is a distance limitation of 50 to 100 km.
AM: Or you’ll lose photons otherwise.
RB: And if you go much longer than that, then you’re not going to get any usable number of photons coming out the other end. They’ll all be absorbed by the glass in the fiber. The situation is slightly better if you go through free space. So, if you go through the air, you can do so well, if you go through the vacuum, even better. So, communicating between a satellite and a ground station on the Earth and a bit of atmosphere at the bottom, then through the vacuum, you are able to close the link and successfully detect some photons, if you are sending from a satellite down to a ground station, or, possibly, vice versa.
AM: That makes sense. So you distribute the keys from the photons from a satellite.
RB: Yes. So the trick here is that you have to use your satellite as a trusted key exchanging node. And if you can trust that satellite as a key exchange node, then you can use this as a way to distribute encryption keys around the globe.
AM: And how far along are you? Because this seems like a very expensive venture to start with. So how far along are you?
RB: Well, SpeQtral is a spin-off company from the National University of Singapore. And in the university, there’s over a decade of research that’s gone into this, developing ground-based free space links and more recently the satellites. We have a couple of satellites in orbit right now. The latest one, SpooQy 1, was put into orbit in June. That’s demonstrating the latest version of our entangled photon source. So that just shows that the source survived inside the satellite. And this technology is space-proved, so it raised the TRL of that technology. The next version of that, which is still a university project, is being prepared for the QKD QubeSat mission. It is a follow-on UK-Singapore intergovernmental mission that will demonstrate full space-to-ground QKD using a CubeSat platform. So, that will launch in 2022 or so.
AM: And that will further increase the TRL of all.
RB: So that will be a demonstration of the full system, and then the technology developed by the NUS side will be licensed by SpeQtral for commercial purposes.
AM: And who will be your clients for that solution once it gets ready?
RB: The QKD CubeSat is an academic project. But the versions that SpeQtral will sell will be for anyone who wants to buy it. These are telecoms operators and banks and financial institutions are our target customers.
AM: Anyone who relies now on …cryptography? …quantum computers? Who is able to use that?
RB: It’s the government, military. Anyone, who has secrets they want to keep safe, should be looking into this, and many of them are.
AM: How difficult will be to integrate this into the existing platforms that you guys have? What would they need to do in order to implement this technology down the road?
RB: It depends. There are many different configurations for secure networks. But this QKD is applicable where you are using symmetric encryption already. So if you are using AES encryption to secure data channels, then QKD can be used to supply the seed keys for those AES links. And these seed keys would normally be supplied over the Internet using public-private encryption, or they might currently be provided by couriers, people with briefcases or USB sticks literally going to each side, plugging these in.
AM: Physically bringing the keys, themselves.
RB: Exactly. So, QKD or satellite QKD provides a more automated, arguably more secure, because you only have to trust the satellite that the link is inherently secured behind quantum physics. It gives you the opportunity to refresh your keys more frequently. If before you were doing it once every six months or so, or once a month, this time you could refresh your keys with every successful flyover of the satellite over the ground station.
AM: At every connection of the ground station with the satellite, basically.
RB: So we were just talking about how to refresh your seed keys of your AES module. There are other ways of applying this. So you can use this for what’s called one-time pad encryption, where the size of the message is the same as the size of the key. So, this is the most secure way you can transfer the message.
AM: That you put in the messages themselves.
RB: Yeah. But that’s very challenging because you can use a lot of keys or very short messages.
AM: Here’s the last one. When do you think quantum computing will potentially be that spread-out that it will start to compromise the existing of security networks?
RB: Well, this is anyone’s guess, really. There’s more than one aspect to this. The goal posts are moving as well. Quantum computers are improving at faster rates than we expected if the reports have to be believed. But also the computing power required to break the current encryption is also going down because people are finding improvements to the algorithms, and whereas before they were might need millions of qubits, maybe that’s halved as the new algorithms…
AM: Let me paraphrase that: will we be on time with the solution?
RB: That’s another aspect of this. And that’s how long do you need your data to be secure? If you need your data to be secure for 50 years, 100 years, then maybe it’s already too late, because if someone has intercepted that data as it travels along a data channel, and they’ve got a copy of that that they’ve stored, they’ve got a copy of your public key, because that’s public, then they just need to sit and wait, and they can decrypt this later, when the quantum computer comes along. And 50 to 100 years we are for sure gonna have a quantum computer. So, for those guys with a 50 year plus requirement on their data security, I’m sorry, it’s probably too late. When it comes to maybe a 10-year lifetime of your data, then you need to think, what is the probability of a quantum computer appearing in that time and what is your risk appetite at that. Maybe, there’s a one percent chance of a quantum computer that’s powerful enough coming along in 5 years. Maybe, there’s a one percent chance that it already exists and no one is talking about it.
AM: You can probably get insurance coverage on that.
RM: So, you have to consider is that one percent manageable or is that already too high? Honestly, we don’t know that really…
AM: Thanks, Robert, thanks for giving us a glimpse of the world of quantum computing and quantum key distribution. It’s really exciting. Good luck with all your projects and wish you all the best.