Marc Kaplan - Quantum Internet: the Short and the Long Term

just introduce our uh first uh first speaker. Our first speaker is Mark Kaplan. He's the co-founder of Veric Cloud and indeed a leading voice in the space of quantum internet. So a warm welcome. Thank you.

Good afternoon everyone. So I'm here to talk about the the effort the worldwide effort or the European effort at least to build a quantum internet. Um I won't really deep dive into quantum things but talk more about the the potential applications of this uh this quantum internet the motivations where we are and so on. But um let's start with introducing you know who we are at very cloud. We work on this topic for a number of years now uh to try to uh to design quantum network applications.

Right? So we know how to build as you will see a quantum networks. The question is what are they good for? And these are a few of the deliverables that we obtained within a mostly research project called the quantum internet alliance uh led by the technical university of Delft. Um so in particular trying to uh to match the theory of what can be done on quantum networks with potential use cases.

This is uh this uh report in the middle. And one of the thing that I strongly encourage you to to look at if you're interested in the topic is this protocol library that we're hosting. I will give the address um the at the end how to reach it. Okay. So quantum networks uh you can read the term in the literature but it covers a lot of different potential uh instantiations depending on the technology that you have in your lab or in your company or to to process quantum information corresponds to different type of networks that you can that you can build right and there are some some networks actually that we can build today and some some stages stages of quantum networks that are very much long-term.

So um today what we have is the the first line here quantum key distribution networks they only require to prepare single cubits send them through let's say an optical fiber and measure them at the end and if you were there uh this morning you could see the experiment uh going on run by never local here and if you weren't there this morning you can visit neverlocal's booth and and see how it works. So it's something that is possible to do even at an industrial level and then you have you know the more so preparing those cubits is easy measuring these cubits are somewhat easy as long as it's just one cubit at a time these are things that we can do uh however um you know beyond that we don't know much for now right so repeating uh you know this notion of repeaters is something that is very hard to achieve for quantum networks. So extending the so photons so fi optical fibers are lossy. You lose a lot of photons half of them every 20 kilometers if I remember well. Uh so um so so you needed repeaters to extend the distance and this is something that is very hard to to uh to build um when you when you can build entangled states whatever that means I won't go into the details here.

uh you know you can also add them to your network quantum memories quantum computers these are the stages of quantum computing and what is interesting for us you know at very cloud and as people are interested in quantum information is that each new stage here you shouldn't wait until you have quantum computers to build quantum networks actually you should make you know start with current existing quantum networks and make them evolve the reason is that each time you add some of the hardware we're here, you will also unlock new applications. And what I want to do uh today is to walk you through this, you know, future history of uh of quantum networks, showing you what is possible today and how what what are those applications that will get unlocked um in the future. Some of them are really weird by the way. Yeah, the goal is not to uh demystify quantum actually. Quantum is mystical and uh I will not try to hide that.

Okay. So what the only thing that is running today at industrial scale on quantum networks is quantum key distribution. I call it the mother of all protocols because although it was not the first quantum communication or the quantum information processing task, it's sort of the the one that captures very naturally uh the the properties of quantum systems and translates them into cyber security properties. So use it. So by by sending these you know quantum states one by one something that we can do very well somewhat easily using photons uh sending them from one party to another Alice and Bob here can establish a key that is unconditionally secure.

What is interesting is that it's something that is provably impossible with classical communication. So using these quantum systems you can do something that is provably impossible with classical uh systems. And by unconditionally secure what we mean here is that if an adversary I mean this property the security derives from the fact that if an adversary tries to observe the communication that is happening because of deep fundamental physics properties he will change the the states and this you can observe you know by by running some statistics over what was uh what was happening. So it's possible to derive a full security proof showing that the key uh sorry the mutual information between the key that Alice and Bob share and the adversary uh is zero you know after some postprocessing it's possible to achieve that quantum key distribution network again I mean not only it's possible today but it's also deployed at a somewhat industrial scale I would say in Europe we have 27 national test beds running at the moment. Uh I mean some countries even have several like test beds running in parallel.

Um in uh in Asia we are at a commercial stage. So it's very hard to know for real what is happening in China but Singapore or Korea have commercial offers if you want to connect to their quantum key distribution networks. Uh in the US they have a slightly different approach. uh I I don't know what's the justification for that but uh but they they could build these networks if they wanted to. Interestingly, I mean, you know, quantum key distribution exists for us.

It's very interesting when you start looking at it as as a building block for further application. And one of the thing that we do at very cloud is we have this application running on top of quantum key distribution networks that transfers you know the security of your communication and the fact that the key cannot be um copied to the security of storage. So we can run distributed storage over several QKD networks um which corresponds exactly to the current situation. There is a small QKD network in Singapore, small QKD network in in Canada and so on and so on. And we can use that to have provably I mean a security that is provably better than using a single QKD network or uh classical networks only.

Okay. So the the problem you know of if you remember my picture earlier right quantum key distribution involves only two parties Alice and Bob of course they want to establish a shared key and this is what happening. So when you want to scale it to a network and have several parties, this is an issue, right? How do you have like how can you have end to end uh end to end uh security when distributing keys over a network uh rather than just pointto-point. Uh this is far far from obvious.

Uh you need to have these repeaters uh to route the the quantum information through the network. And at the moment there are two potential solutions uh to this problem but these are lab experiments. The first one is to distribute entanglement of the on the network uh and the second one is to use quantum error correction and they're both very very hard uh to achieve. It's running so entanglement distribution works in lab. it starts to be uh deployed in a in significant or relevant environment.

Uh at this stage of repeater network there is also an interesting story that you can use quantum information to have more secure confidential computing classical confidential computing. So you can run. So my example here is that you have Alice and Bob that um try to establish you know the intersection of their uh tables for example and uh it turns out that you can prove that there is an an advantage in terms of the security assumptions that you need to achieve this task on quantum networks compared to classical network. Of course it's a task that has numerous applications. Um, I'm not going to oversell it and tell you that this is where quantum networks are relevant.

But at least it's something that should be tested and as far as I can tell, it has never been tested out of a research lab. And it requires exactly the same hardware as quantum key distribution. It's just that if you want to reach long distances, once again, you need these quantum computers that we don't know how to how to build today or that we don't know how to build in a real environment. Okay, now let's get even weirder. Um, unclonable cryptography.

Now suppose now I have quantum memories. Quantum memories are devices that can receive the quantum states that are know flowing through the the network can store them for an unlimited amount of time and when I need I can either release those quantum states and send them further to another party on the network or measure them. So there's no processing there's no processing in the memory it just stores the state and send them. It turns out that this stage in my opinion is one of the most interesting intermediate stages uh that we have in in the building of these quantum networks um because it unlocks all these tasks that we usually put under the name of unclonable cryptography. So quantum money protocols we'll hear about them more I think today uh tomorrow.

Uh and clonable encryption it can store a can store some information on the server with a physical guarantee that the server cannot copy it on another server. So cannot share my information. Uh certified delion I can ask the server to give me a mathematical proof that it has erased the data that I uh that I stored. This is very interesting and interestingly quantum memory is now uh there are a few startups that start commercializing quantum memories. So I believe that we'll see more and more experiments in the future.

And by the way um you know even though there are startups it's still very early work and we're starting a project with never local to try to uh determine what are the parameters that we need for these quantum memories to show anything relevant for applying quantum money protocols in the blockchain environment. See so this is how we relate to uh to blockchain in general. We know that they could be useful and it's worth um investigating it. But it all starts with simulating and you know spending time to uh to implement the protocols first on simulator and then hopefully on real hardware. Uh last but not least this is of course I mean this is this is the killer app.

This is the killer app for quantum networks but also it's the furthest because you need quantum computers. Once you have quantum computers with the devices that we have today, the ones for quantum key distribution, you can run secure quantum cloud computing. What do I mean by quantum cloud computing is that these devices can send an encrypted version of the the circuit the quantum circuit to be executed on a on a on a on a distant quantum computer. And uh by secure execution what I mean is that it's the server remains blind. Everything is encrypted and it's also verifiable.

By verifiable what I mean is that the client can verify that the computation that the server is doing is the one that it has paid for. And this relates to very deep questions again in um in physics. Um it's related to fullomorphic encryption but also a little bit different. This is all I want to say here. We uh we also participate in this type of uh say research uh problem here and we showed that the the hardware by the way yeah we also manufacture hardware for quantum communication and we showed that our hardware is well suited not only to run blind quantum computing but also to scale it to a multi-user uh case and although we didn't have a quantum computer we could use mini quantum computers uh here at University Sapiens in Europe.

Okay. So in conclusion, uh quantum communication is today a mature technology. Uh you can uh purchase hardware to uh to do it and we've seen it this morning. But it only runs a single applications which is quantum key distribution or let's say there is very nice work in progress to try to push further in deploying more and more application in today's hardware. Designing use cases is super interesting but it's also super hard.

You need smart people to design even if we you know what you can do on your quantum network applying it to a use case is requires a lot of work. uh and and this you know long-term horizon we know it's a killer app for for quantum computing where that you can have privacy by design on quantum networks. Uh it's this long term that we call this quantum. Thanks a lot. Uh this is the address if you want to look at and learn more about uh quantum communication protocols here.

Uh and I strongly encourage you to start here if you're interested in the topic. Thank you. [Applause]