Data Engineering Podcast

Share on social media:

Support Data Engineering Podcast

[00:00:14] Unknown:

Hello, and welcome to the Data Engineering Podcast, the show about modern data management.   Have you ever woken up to a crisis because a number on a dashboard is broken and no 1 knows why? Or sent out frustrating Slack messages trying to find the right dataset? Or tried to understand what a column name means?   Our friends at Outland started out as a data team themselves and faced all this collaboration chaos.   They started building Outland as an internal tool for themselves.   Outland is a collaborative workspace for data driven teams like GitHub for engineering or Figma for design teams.   By acting as a virtual hub for data assets ranging from tables and dashboards to SQL snippets and code, Atlan enables teams to create single source of truth for all of their data assets   and collaborate across the modern data stack through deep integrations with tools like Snowflake, Slack, Looker, and more.  

Go to dataengineeringpodcast.com/outland   today. That's a t l a n, and sign up for a free trial.   If you're a data engineering podcast listener, you get credits worth   $3, 000 on an annual subscription.   With our managed Kubernetes platform, it's now even easier to deploy and scale your workflows or try out the latest Helm charts from tools like Pulsar, Packaderm, and Dagster.   With simple pricing, fast networking, object storage, and worldwide data centers, you've got everything you need to run a bulletproof data platform.   Go to data engineering podcast.com/linode   today. That's   l I n o d e, and get a $100 credit to try out a Kubernetes cluster of your own. And don't forget to thank them for their continued support of this show. Your host is Tobias Macy. And today, I'm interviewing Pranayha Narong about her work at Aliro building quantum networking technologies and how it impacts the capabilities of data systems. So, Pruneja, can you start by introducing yourself?  



[00:02:09] Unknown:

Thank you so much for having me here, Fais. My name is Pruneja Hanar, and I'm on the faculty at Harvard. I also   simultaneously   wear a   CTO of Alira Quantum,   where we're building scalable quantum networks, and I'm really, really excited to tell you about what these are and how they might be able to to make our lives better. And do you remember how you first got involved in working with data management or at least in areas that are adjacent to that   space? Absolutely. So I was a graduate student at Caltech   and, you know, just started thinking about how GPUs can have an impact in   predicting various quantum phenomena, and, you know, this is before   everything was GPU enabled and   a while before people were talking about GPUs being a daily computational life.  

And that's where I really got excited about new computing architectures and have an impact on problems that are independently of interest, and that's been a theme   throughout my career, throughout my my time as a faculty member as well, where we've been thinking about, you know, if you had the biggest, baddest computer,  

[00:03:12] Unknown:

classical quantum combination thereof, how would you use it to answer the the most important problems facing us today? That brings us now to the work that you're doing at Aliro. I'm wondering if you can share a bit about what it is that you're building there and some of the story behind how it came to be and why you decided to try and productize this and focus so much of your time and energy on it. Yeah. Absolutely. So, you know, we started out,  

[00:03:36] Unknown:

my lab and a couple of students me and we started talking about what does it take to   bring quantum technologies, quantum computing, and other aspects of quantum technologies, and we'll dive into that in in a few minutes, to everyone. What does that mean? Right? And this is now a few years back, you know, before some of the big announcements in cloud computing that we've seen,   and it became obvious that quantum devices in their   current form, even if you take everyone's road map to to be as canon,   turns out that getting to scale will involve connecting these devices.   And, you know, that's just the reality of classical computing.  

We don't expect it to be any different for quantum computing for even more profound reasons than the reasons in classical computer.   So we started thinking about, well, what does it mean to connect devices? And, you know, should we think about it purely from,   a computing standpoint, or is it more general than that? And that's when my team got into, you know, thinking about cloud networks, in particular, cloud networks that are scalable   or universal.   And   the idea with these networks is that, you know, it's not specific to   connecting 2 endpoints or 3 points for secure communication.  

It's not specific just to connecting to superconducting   computing devices.   It's more general than that. It's a network layer that can actually work with various different types of hardware, can work with different types of use cases,   and that suddenly starts to sound like the early days of of the Internet where we were starting to talk about, you know, connecting devices.   Yeah. That's how we came to this field. It's a little bit different from, you know, where   most people you meet in the field have come primarily from the cryptography   or the secure communication side. We came at it from the standpoint of, you know, what are general network pieces that will enable both computing and secure communication.  

As we were doing that, we also realized the 3rd part of this triad is, of course, distributed quantum sensing, and and the 3 go and they can.  

[00:05:35] Unknown:

And in terms of the initial use cases that you're targeting, I'm wondering how you went about that sort of discovery and selection process to understand   what is the inroad to be able to say, I've got some functional   capabilities. I want to start figuring out what is the product market fit, how do I determine who are the early design partners to be able to actually start putting this into production   and stress testing the, know, theoretical ideas in a real world environment?  

[00:06:04] Unknown:

The big hope here is a general quantum Internet. This is, of course, going to transform our lives. You probably heard this spiel from a couple of different people,   and you start to look at, well, what is needed in order to actually   make such quantum Internet quantum architecture at scale   a reality? And it turns out we have some pieces of it today. Right? So we have some devices that look like endpoints.   We have some pieces of it that look like a network, but we don't have everything.   1 of the biggest things we're missing, of course, is a quantum repeater. Now before I dive into what's a quantum repeater, let me just take a step back and say that, you know, when we think about networks, right, and when we think about cloud networks, we're really thinking about entanglement generating and using networks, And this is fundamentally different from some of the work you might have heard about from people in key distribution, for example, which, by the way, is a field that's been around for some time.   So when you think about what it would take to get to larger distances beyond metropolitan distances   to actually have multiple   points, a complex apology in this network, what it takes to include both terrestrial and satellite based pieces, suddenly you realize you need a lot of new quantum networking components.  

And some of these are commercially available now, and some are commercially available perhaps from just a single company.   And these include things like quantum memories, quantum transducers, quantum frequency conversion.   The list goes on. So our   proposal, our idea   was, let's start with a turnkey integrated prototype network.   And   in this prototype network, we take the best pieces you can get today,   and   we show you what this looks like   as as a functioning network. Right? So you turn it on,   add something goes. Now it's not gonna be the speeds you want. It's not going to be all the use cases functionality you want,   but it will allow our customers to actually build the confidence they need in order to give us access to the fiber they have in the ground or even to build a dedicated network for them. Because it's a big ask to tell people, you know, we wanna just rip out everything you're doing, get networking today, and replace it. Or, you know, there are some ideas in how you could have both a quantum and a classical channel. There'd be some codesign. There there are interesting design questions to be answered.  

So building these early networks, and it's you know, pick your favorite analogy, ARPANET,   you know, early days of Cisco. I mean, there's so many analogies here today,   whichever 1 resonates with you. But the important piece is that it's towards that connected   on an Internet, and this is where you can imagine having not just secure communication.   You also can imagine having your favorite   compute devices on a cloud. This is perhaps the only way that you could get to blind quantum computing. This is where you wanna make sure that people can't actually see what you're computing. It's something we take for granted in other contexts.   And, also, distributed quantum sensing. Right? So there's this idea. Well, they're going to be quantum devices. They're going to provide us higher sensitivity,   more information. That's great. It can all be connected in this, quantum cloud. So we start to engage customers across that entire spectrum. I basically, Peter told you, in tech terms, all the people that we're working with. So there are folks who are interested in secure communications today the various parts of defense and intelligence communities, for example, various big corporations that have important data to protect. On the computing side, we have both classical and quantum computing folks who are interested.  

I, you know, don't wanna   take this time to drop any names, but just to say that, you know, there are various architectures that have been proposed, trapped ions, superconducting,   but they're purely photonic architectures.   And then there are also ideas in how you could actually combine   a couple of these into this best of both worlds steps scenario. So use 1 as memory that goes with something that has faster gates being being a company example. So 1 of the things that, you know, when we start talking about this very, very complicated architecture, it's it has all this going on is,   so are you going to go from concept to build on day 0? And the answer is no. In fact, like classical networks, we need to do a lot of simulation   and direct emulation of various pieces before we get to the prototype.  

And within the Alero team, we've, you know, built up a in house simulation platform.   It's our in house sandbox.   We also have now an emulation platform, as you can imagine, with various quantum pieces. You know, each of these components, right, I mentioned a minute or so ago, has various parameters it needs to hit for this entire architecture to go. And so now it's a range of parameters. You can think of there being a minimum performance parameter, and then there's something that is perhaps more optimal that allows us to relax constraints in some other part of the system. And that kind of systems engineering is what we're doing   within the both the simulation and emulation piece while building these prototype networks.  



[00:10:57] Unknown:

A lot of different directions that I'm interested in digging into this. And 1 of the first things that I'm curious about is   given the fact that you're are building this fully integrated   system of various components, it's definitely very analogous   to distributed systems problems that exist in terms of data management, data processing.   And so that brings up the question of   consistency and availability guarantees in the network layer and how you're able to   model that given the fact that you're dealing with probabilistic states and not concrete states of, you know, on or off you have to deal with? What are the gradations of probability that any given piece of information is propagated to a certain point in the network?  

[00:11:37] Unknown:

Yeah. So, actually, you know, when you think about   the underlying technology here, in part, what makes the the Aliro network,   in a way, universal is that it's an entanglement generating network. So it runs this nice link layer protocol   that that generates elementary   point to point entanglements,   and then it runs these network layer protocols that do the swapping, entanglement, purification. Now some of the things that we are talking about here with the layering, the layer structure   has direct analogies to classical networking, but some of the pieces are, you know, different.   And what we're trying to actually   be cognizant of is that, you know, the way you think about it in customer networking might not directly pour over. There might be some differences. Of course, we don't wanna reinvent things that people know, but we also don't wanna force and mold onto quantum networking if there is a more natural   way to do it. So, you know, you can think of I mentioned the word swapping and purification. So that's happening in the network layer, and the swapping here is going to pull together   various different elements or entanglements together to get end to end entanglement that you're then delivering to the various quantum applications. So entanglement really is the resource that you're distributing across this network.  

Now if you think of it as just the best entanglement I can generate, that is frequently insufficient, and this is here towards the probabilistic component that you point out. So you need to actually think about improving the fidelity of the entanglement,   and   that involves doing some form of purification or distillation.   And there are various protocols that exist for this, some which I put out some web codeveloped with the Liro, and, of course, there are others into the community that that exist. And you can think of these as trying to meet some quality of service   requirements of the quantum application. Right? So there's some amount of entanglement, and this is the worst way of thinking about entanglement because it's a very quantum resource, and I'm describing it as this very classical resource that some amount is going to get everywhere.  

But you need to hit certain thresholds for the application to actually go,   and all of that is being handled, in our case, in the network layer. Now   this is   in contrast with some of the work that's happening with, you know, what people are doing a key distribution based Clarm network. So I contrasted entanglement generating using networks, the fact that the universal networks with these, you know, current QKD networks, because they're based on this prepare and measure quantum technology. And,   realistically, what that means is that those networks can only ever really be used for key distribution because this idea of   sharing end to end entanglement and using it as a resource for quantum applications isn't something that you could do with this network.  

And, yeah, as you very correctly pointed out, it's still, nicely happening in the the network layer. Another interesting element of this is  

[00:14:21] Unknown:

the   junction points between classical and quantum networks, and the obvious analogy is   the junctions between fiber and copper networks that we have been going through the process of upgrading. And I'm wondering   what does that look like in terms of the actual   handoff and if there are any   protocol translations that need to happen because maybe you're not using TCP over a quantum network, for instance.   And what are the   architectural locations where it makes sense to put in those junctions where you're going to say, this is the point where I'm bridging from my classical to my quantum network. So does it go into an ISP? Does it go into the data center? You know, what do those   design questions look like as you're starting to play with these network topologies?  



[00:15:05] Unknown:

It's very early days at the moment. So, you know, what you can send over the quantum channel is very   limited. And right now, you wanna do as little as possible over the quantum channel and as much as you can over   the classical channel. And that's just because, you know, we don't have repeaters, or the repeaters that exist that are embarrassingly   slow. You know, we're long ways from from, you know, sharing, and people always ask me, well, will I be doing a a Zoom call over a cloud network? And   long ways from that, in fact, that itself might not be something that makes sense. So we're not replacing the classical network here per se, but actually   combining pieces of it, doing things over the classical and the quantum channel somewhat simultaneously.  

Now protocols need to be developed, and that's actually where a lot of ideas, including SDN,   are crossing over from classical to quantum networks. I know classical networking, people are already, you know, writing,   but quantum networking, we're still learning about SCN and thinking about what that means for quantum networks and how we can best use it.   And this is where, you know, there's the technology junction, but there's also the people junction that we're making and slightly different from the point,   you perhaps asked about to us. But I just wanna say, you know, we have noticed that people who are coming from a classical networking background   to cloud networking are really really bringing something very valuable.  

It's preventing us from making sort of the same mistakes or perhaps going down the same rabbit holes that people would have in the early days of classical networks. And, this is where, you know, being Boston based has been great.   We've been able to get   top networking   talent here as well as quantum talent. Some other classical analogies that come to mind as we think about quantum network as well. Do we need something that looks like a quantum bus? Is there some fabric functionality?   Is it all going to be, you know, what's a Mellanox   of cloud networking? How do I think about a bus? And I I think all of these components are being developed, some by us at Alero,   and how they're going to be integrated into this, you know, network that's going to be general purpose   is still being figured out. So that's,   that's where we are with with these technology choices right now relative to, you know, cost of networking.  



[00:17:29] Unknown:

Another thing when you're talking about network technologies is what are the ways that it's going to fail? So with classical networking, you can hit TCP congestion issues. You can have issues with BGP routes being wrong. You can have problems with, you know, some undersea fiber cable gets cut so you can't reach another continent for some reason.   And   all of the different   failure   mitigation systems that have been put in place as far as, you know, traffic rerouting,   TCP retries, all of these different things. And I'm wondering   how that maps into quantum networks and some of the ways that you have been simulating those failure modes to be able to design for those eventualities.  



[00:18:07] Unknown:

Excellent question. Actually, some of the failure modes that you described are common to classical and climate networking. Right? So, historically, if you think about classical networks, and not that you say historically, like, they're gone. No. They still exist. You simulate and you model what that abstract behavior of this network looks like, and you emulate to see what the functional behavior looks like and you see some com complex system behaviors.   Now we can simulate   many, but not all pieces of a quantum network, and this   is partly where   things get tricky because all of these quantum components that I pointed out, explicitly the repeater, it's very, very hard for us to directly stimulate. Right? It's a quantum problem in itself, so that right there becomes a challenge. But I can better emulate it   by, you know, asking somebody for an actual repeater that I, together with the simulation, put into an emulation platform. And that's something we've been doing.  

This actually, you know,   allows us to think about simple problems like connectivity and queuing complex problems like you pointed out, like congestion, thinking about new protocol design, fault management.   And   identifying network component failures is gonna be very hard for, you know, various cloud technology just because so much of it is new, and there are, you know, very few things that we know about how it operates in the field. So we're trying to derisk all of that by figuring this out while we build these prototype networks.  

[00:19:30] Unknown:

As far as what it means for people who are working in the data space,   What are some of the new capabilities that quantum networks can enable? What are some of the ways that   quantum computing and AI running on those systems is going to change the way that we think about data distribution and data modeling,   and what are some of the ways that it's going to stay the same regardless of which network technology you're relying  

[00:19:56] Unknown:

on? So this is a little bit of a convoluted answer, because how such networks will be useful to people   working in various areas of data science   in the near term goes through the route of quantum computing, so stay with me. Right? And so it turns out that, you know, existing quantum computers, they're they've been announced. People can do toy problems on these, but the valuable problems in   conduct optimization,   the kinds of algorithms that have been proposed for quantum machine learning, for example,   those are not accessible by current quantum computers in their present form.  

Think to not gotten to a stage where, you know, you're seeing this practical quantum advantage. I didn't say logistics. You know, people ask about that a lot because supply chain has been on everyone's mind, but there are other problems that roughly fall in that same category that you really can't access with cloud computers today.   However, what you can do is go through a scale up approach, which goes through the route of connecting these cloud devices.   Right? So if you think of, like, a discrete GPU,   now you can interconnect   many of these using the kind of quantum networks that we're building. And this is why it's universal quantum network. It's the same thing we're building is relevant for scaling up quantum computers, and and it's the same technology for actually doing secure communication.  

So if I use the classical analogy here, it's going to be necessary to build a multi core, a multi processor quantum computer. Right? And this is all so obvious when I say it because people say, yeah, of course, great. But this is where, you know, you have various quantum cores processors that connected using a quantum bus,   and, you know, you have some multicore TPUs that connected using a quantum fabric.   Now because we have a universal quantum network, our quantum network can provide the required quantum bus, this quantum fabric functionality,   essentially generating these entangled qubit pairs between 2 QPUs in the cluster.  

You can basically integrate, and you can think about implementing   distributed quantum gate circuits and do the type of problems. Right? This is ultimately getting to your question to ask, how do I do 1 of these problems that are of interest to people, you know, in AI today? And the path to that really goes through   scaling up with these quantum algorithms distributed on computers. So that's where I think there is a lot to be done in the near term and where there's going to be a direct intersection between people in data science and people in areas of cloud networking and cloud technology.  

[00:22:18] Unknown:

Modern data teams are dealing with a lot of complexity in their data pipelines and analytical code.   Monitoring data quality, tracing incidents, and testing changes can be daunting and often takes hours to days or even weeks.   By the time errors have made their way into production, it's often too late and the damage is done.   DataFold built automated regression testing to help data and analytics engineers deal with data quality in their pull requests.   DataFold shows how a change in SQL code affects your data, both on a statistical level and down to individual rows and values before it gets merged to production.   No more shipping and praying. You can now know exactly what will change in your database.  

DataFold integrates with all major data warehouses as well as frameworks such as airflow and DBT and seamlessly plugs into CI workflows.   Visitdataengineeringpodcast.com/datafold   today to book a demo with DataFold.   As far as the actual   sort of capabilities of quantum networking, does it break the faster than light problem that we have with classical networks because of the fact that you're dealing with entangled pairs? How do you manage that? You know, what are the expectations that we wanna set for people who are thinking about quantum networking and all of the magical, you know, frolicking unicorns that it's going to bring to our data problems?  

[00:23:36] Unknown:

Okay. So we never go faster than speed of light. This this paradox, you know, thinking about it, tag it, and teleportation,   you're also at the speed of light. This has been resolved.   Everything we're doing is well within the known laws of physics, and that's a good thing. In terms of where things are and when it's going to start to have an impact, right, prediction is hard, especially about the future, and this is a moment where it becomes true.   So what I can say is that   connecting quantum computers,   connecting quantum devices, these small networks, this is happening right now. Okay? There are big networking companies,   big telecom giants that are looking at quantum networking technology today. That's happening. Now does that mean it gets to the end user in the next 6 months or a year? I don't know. I think it's gonna take a little longer. Depends on who you define as an end user of a quantum network. After they're thinking of, you know, quantum computers connecting quantum computers because they're trying to be the first to solve a very hard problem Yeah, you'll start to see that in the next 6 months a year because various quantum computing companies are talking about using such a fabric, you know   distributed computing architecture using our kind of networks. So   but   You're not going to be you know, asking for entanglement as a service to arrive at your home anytime in the next near future at least   And that's because, you know, we have various pieces in this technology that need to be figured out, especially as we get towards that quantum Internet. And I think some of this has to really be   in those network components that I pointed out. We have them in a functioning form today. They are far from optimal. So so we need investment in this area.  



[00:25:20] Unknown:

Are there any sort of notable differences in terms of bandwidth capacity or overall end to end latencies that are available because we're using quantum networking? You know, what are the limitations look like where people are used to looking at maybe 1 gig, 10 gig, 100 gig connections for, you know, the different fiber interconnects?  

[00:25:39] Unknown:

Oh, I think, at the moment, we need a real throwback here. You know, think late eighties   in terms of, you know, speed, in terms   of latency,   in terms of, you know, what you're expecting   from the network.   However, and this is the upside of doing it the second time, we're not going to go through the same slow technology progression that we did with the first wave of the Internet.  

[00:26:01] Unknown:

For people who are now looking at using quantum networking to be able to bridge to their quantum processing units for doing quantum enabled AI,   how does that change   the overall approach to thinking about the ways that they're   structuring their delivery pipelines, their data modeling?   Are they going to be   locating storage directly adjacent to these QPUs,   or are you going to be feeding reduced datasets into QPUs so that you're not   trying to replicate your entire data store? Like, how does that shift the way that we think about our overall workflows?  

[00:26:38] Unknown:

Yeah. I think people are talking about and not just talking about building   quantum data centers already,   and this is, you know, towards   there's some proximity choices that people are making implicitly and explicitly.   Right? So I think that plays into   the question you just asked.   I think in terms of really modifying workflows   and, you know,   making major changes to how people are doing things in in AI right now, I'd say it's, you know, not something that you should do on your next build cycle,   but it's something that   you should start looking at and thinking about, is there   going to be a cloud data center that is within x miles of where we're located? How would we integrate? Would we need to integrate? And I think those types of questions really should be asked.  

And, in fact, they are being discussed, you know, at various levels in the government. They're being discussed, of course, at various levels, big corporations.   I hope that that information makes its way   out to folks maybe who don't have an internal r and d arm that's looking at this area.  

[00:27:44] Unknown:

Other aspects of the networking layer of the data stack are things related to security, which you touched on a little bit as far as   because you're working with   entangled pairs, you don't necessarily have to worry about the encryption protocol because   it's entirely private due to the interconnect that you're building. But how do you think about sort of data privacy considerations   in that layer?   How do you map the handoff from the classical to the quantum network to manage some of those questions of data privacy at the different stages? And, you know, maybe at the bridge, you say, I'm actually going to strip out the   encryption because it is adding to the overhead of this transmission. So we can actually optimize the transmission from the quantum bridge into the quantum computing unit and then being able to re encrypt on the reverse journey.  



[00:28:33] Unknown:

There are various vulnerabilities of the network, some that are physical, some that are at the handoff, and some that, of course, are the network layer. So you get physics based security in quantum networks. That's something that everyone hears about, everyone talks about. I think what's harder is when you're, you know, using both the quantum and the classical channel   to figure out all these other   side channel attacks and figure out all these other vulnerabilities. And we're working with, you know, customers to identify   problems that are specific to how they're going to deploy these networks and how we're going to, you know, protect against attacks like that, especially if interest or intelligence community customers.  



[00:29:11] Unknown:

And so this is definitely a very   interesting and nascent field that has a lot of potential for future development. And   a couple of questions related to that are, what are some of the time frames that we're looking at for actual practical adoption of these capabilities   and for people who want to get involved with this space as either implementers or early testers of these capabilities, what are some of the useful onramps to that?  

[00:29:39] Unknown:

Excellent question. And, you know, I wanna point out that there are various companies that are hiring in this field already,   and we've only seen that hiring ramp up over the past year. And based on where   folks are headed in computing, quantum networking, I think the demand is really growing. So   for folks who are on the classical side,   I encourage you to think about this at the moment to make the jump. Few reasons. 1, of course, as I was pointing out, that   the background people have in classical networking   is actually going to be important towards enabling these quantum networks. And towards that, various conferences in classical networking have started to do various   small   your your first day of cloud networking type sessions,   organizing a couple of things with IEEE   to actually, you know, get some some nice pedagogical review articles out. There's been a explosion of people who've said, hey. Is there a good textbook? I know there are efforts to get   accessible   cloud networking,   you know, digital textbooks out that people can actually start on and look at some of these interesting problems,   see them framed in a way that is reminiscent of what they've seen in in cost   previously. This is not an area where we've started to have the kind   of open source   lectures that you see on the cloud computing side, but it's come. So the second thing I wanna point out, you know, especially to people who   are graduating   or, you know, going back to their masters, think about taking,   you know,   a class that roughly touches on quantum because I think a lot of the time that people spend in making the jump is spent on you know, jargon, figuring out some of the notation,   and that's all it is, really. It's notation and jargon. It's conceptually   not particularly   challenging to follow for people who have a stem background. It can really make you very valuable to companies looking for people in kind of networking and quantum computing,   especially software developers, particularly, you know, people who have a background in large infrastructure thinking about distributed computing. There's a large demand for those folks on the quantum computing side, and it's not going to be satisfied by the number of people we train in, say, theoretical   cloud information.  



[00:31:49] Unknown:

In terms of your   work on building a Liro and experimenting with these early quantum networking technologies and interacting with design partners, what are some of the most interesting or innovative or unexpected ways that you have seen these quantum technologies used either for just networking in general or for data systems in particular?  

[00:32:09] Unknown:

I've been pretty surprised by how people were doing time sensing,   long baseline astronomy, had started to actually   connect these and use networks to   connect various sensors and do experiments that you would not necessarily   think of, you know, without that quantum enabled system. So that was pretty exciting. I can't think of a company that's doing that off the top of my head, but, you know, I know of various research groups that are trying that. I've been pretty surprised by that. That was a cool use of cloud network technologies. Yeah. It's not really data science related in that context, but but it's generating really, really cool data for our astrophysics friends to   analyze and maybe   find, new stuff.  



[00:32:51] Unknown:

Absolutely.   And   I guess as a brief diversion,   I'm curious what   kinds of capabilities quantum networking   enables   as far as things like space exploration   or   just sort of spatial communication?  

[00:33:09] Unknown:

So   turns out there is such a thing as quantum enabled communication.   For example, for a dedicated moon or Mars link. Actually, I put out a peer review paper together with my colleague on this exact topic. It's of interest to NASA, and part of it is   overcoming some of the detection limits from classical   Tetris. So you're, you know, encrypting and then using this joint detection scheme that we propose. So there are applications, actually, of such kind of protocols and, you know, these dedicated space communication links. Another   related area I see where people are talking about, you know, using the satellite, creating this fabric. Right? So that's good. But if you could do that and share   confirmation across it, then it's even better.  



[00:33:55] Unknown:

And I think that that's an application we'll start to see, you know, various satellite based startups explore in the near term. Yeah. Definitely lots of interesting things to dig into. And in terms of your own experience of working in this field and starting Aliro and working with some of your design partners,   what are some of the most interesting or unexpected or challenging lessons that you've learned in the process?  

[00:34:16] Unknown:

Oh, gosh.   This is a long list, but let me pick maybe the top 2. I think the first has been that hardware just takes longer to arrive, and this was before supply chain issues. So we just learned that the hard way, that making stuff takes time and it never happens   on time. So, you know, working with some of our component vendors, we realized that we really need to, you know, build in plenty of buffer,   especially because of the, you know, level of specialization   of these components.   And now with trying to navigate   not just supply chain issues, but also supply chain security, you know, we we really need to   allocate a lot of time for that.  

And, initially, we thought it's just us. So that was, you know, part of the lesson we learned is that, no. It's not just us. In fact, every company out there is is facing this. And the sooner we get that information   out to our partners and customers, the better off everyone is. So that's something that we and others learned, and now we've started to, you know, have,   share Slack channels and discussions around, you know, getting specialized parts from particular   other vendors. Another lesson that, you know, I think has been interesting   for us is to really validate the claims that are coming out from folks that don't have peer reviewed publications to back it. I think that was, you know, another 1 that I personally learned. I think coming from an academic background, I always thought, well, if somebody's reporting on it, surely, there must be some amazing paper or IP behind it. And in many cases, that was true, but there were some cases where that wasn't true, and I was I'm pleasantly surprised because, you know, I would look at it and say, oh gosh. They've done everything that we wanted to do. I have to wonder how they've done it, how they figured this out so quickly,   and that led to, you know, me then realizing, oh, wait. The way they're saying they've done it, it's actually not making sense. So, yeah, those have been a couple of, you know, lessons for me coming from an academic background   to this very   fast moving tech field and and lessons that I've learned personally.  



[00:36:21] Unknown:

For people who are   interested or   curious about the capabilities   that these quantum technologies and quantum networking protocols enable, what are the cases where it's the wrong choice and they should just stick with classical networking   and not wait around for the promised future of quantum networking capabilities?  

[00:36:41] Unknown:

You know, I think the consumer case is something that I pointed out. Right? If you're looking for   your favorite service provider to offer   as a service in the next 18 months. Or if your service provider is doing that already,   you really have to think about, you know, am I using it for a quantum application, or is this a cool fact? As these devices and networks become   more generally embraced, I think you will see a transition,   and that'll be, you know, the right time to start to embrace these as a personal consumer.  

[00:37:12] Unknown:

As you continue to work in the space and push forward the technologies   and the capabilities   for quantum   technologies and quantum networking at Aliro? What are some of the things you have planned for the near to medium term or particular areas of research that you're excited to dig into further?  

[00:37:28] Unknown:

We're really excited about repeaters. We're really excited about various repeater technologies.   So, you know, whether it's going to be solid state,   whether it's going to be these atomic memories, then going to a repeater,   whether it's going to be a totally different repeater architecture or photonic that reveals itself, That's that's something we've been very excited about because that is the key limiting factor  

[00:37:50] Unknown:

in the speed of these networks at the moment. For anybody who wants to get in touch with you and follow along with the work that you're doing, I'll have you add your preferred contact information to the show notes. And as a final question,   I'd like to get your perspective on what you see as being the biggest gap in the tooling or technology that's available for data management today. You know, I think there are a lot of gaps. I'm going to give the quantum answer. That's what I experienced.  

[00:38:12] Unknown:

I think there's a gap in the various quantum tools that exist to write programs, test things out without   access to all different types of hardware that exist. Right? So, you know, you can get access to superconducting computers relatively easily via IBM or these other   big providers.   But some of the other   types of qubits,   tracked ion or all photonic,   the access has been limited for individual users. And sometimes you just have a question like, hey. I just wanna test this out,   and I don't actually wanna run a high value calculation, or do I wanna set up a calculation   that a collaboration to do 1 calculation?  

It'd be really helpful to have tools that, you know, allow you to do that, whether they're   really, you know, fancy simulators running in the background or instead of calculations somebody's run and just made available   to addressing these questions. But, yeah, tools like that exist, we'd be very, very happy. I would be happy because I had these, like, weird things I wanna run sometimes, and I just can't do it without this   high barrier.  

[00:39:11] Unknown:

Alright. Well, thank you very much for taking the time today to join me and share the work that you're doing at Aliro and some of the research that you've done in the space to enable these quantum networking technologies. It's definitely a very fascinating field and 1 that I'm excited to see come to fruition. So I appreciate all of the time and energy that you and your team are putting into that, and I hope you enjoy the rest of your day. Thank you.   For listening. Don't forget to check out our other show, podcast.init@pythonpodcast.com   to learn about the Python language, its community, and the innovative ways it is being used.  

And visit the site at dataengineeringpodcast.com   to subscribe to the show, sign up for the mailing list, and read the show notes. If you've learned something or tried out a project from the show, then tell us about it. Email hosts at data engineering podcast.com   with your story. And to help other people find the show, please leave a review on Itunes and tell your friends and coworkers.