- What the 3 phases of implementing a quantum network are
- What happens in each stage and how each phase builds on the previous one
- What support your organization will need in each phase of implementation
Why is it important to have a framework of modes or phases for implementing a quantum network?
We think it's so important to have this three mode framework because quantum networks are such a nebulous concept. It can be very daunting for organizations to even get started thinking about how they want to approach this, and because of the big promise of quantum technology it's also very tempting to dump a lot of money in and just go straight to full-scale deployment. That's not the most efficient and effective way to unlock the full potential of this technology.
We propose a three phase framework that can be applied in all kinds of industries.
In brief, what are the phases, or modes, of the framework for implementing a quantum network?
So first, we have design and emulation mode, where you create a plan. It's a very involved process.
Then there's pilot and trial mode where you create a small testbed of sorts, try things out, debug and see what you need to change before you scale up.
And finally full-scale deployment mode, which you've already laid a strong foundation for in the design & emulation phase and the pilot & trial phase, and you can confidently invest larger amounts of money and achieve more ambitious goals to unlock the value of your quantum network.
What does the design & emulation phase entail?
Emulation mode is all about emulating, designing, and validating your quantum network.
Before building a network of any scale, it's necessary to first identify the plans, goals, budget and risk, then you can start design.
Design involves choosing and optimizing hardware configurations, protocols, and more. There are so many pieces in this puzzle, and you need to approach each with caution and care. You also need to think about how these pieces will fit together because there are a lot of permutations, and they all affect each other.
This step ensures that you plan with a data-driven-and-optimized realistic vision.
Some key steps within this process of emulation and design include:
- Assessing what is required in a quantum network implementation to meet the specific requirements of your organization. So for example, if you are a three letter government agency that is investing money in this to provide security, tactical advantage, whatever it might be, your specific requirements are going to be very different from a telecom company that is investing in this as a business to provide a service for users to pay for. The things that you prioritize, and the order in which you do them will change a lot based on this, so it's really important to start by defining these in a concrete way.
- Creating realistic performance estimates for concrete network design drafts and plugging these into a quantum network simulator to see if they will provide the intended value and support your applications. There are a lot of different quantum networking simulators out there. At Aliro, we decided to make our own because we needed it for our own internal work, designing networks, working with customers, and also designing protocols. We decided that it was worth making our own because there were no other quantum network simulators available that provided everything that we needed for this process.
To plan the whole system, you need to model the whole stack, and you need flexibility in how granular your model is. So that's why we created our own Aliro simulator. It does exactly that really well, and it has streamlined our whole process of design. - Once you have a simulator like this, and you're running your designs through it and seeing how they perform, you can start to make informed decisions about the necessary trade offs in cost, performance, and node placement. For example:
- If you want your network to perform well, it's going to cost more.
- If you want your nodes to be in specific locations (like specific cities or important facilities) there might be a trade off in having a node in a specific place versus having a node in a more optimal place.
Without a simulator and this data-driven process, it's very hard to make these decisions and weigh trade offs.
- Evaluate the performance and interoperability of competing hardware modalities and devices before purchasing them. Sometimes you want to choose one over another because it's more affordable. Sometimes you'll choose another because of a specific aspect of its performance. Often you'll need to choose a certain subset because they're the ones that are capable of working together. This can be something as simple as that they operate within the same wavelength ranges, or even more involved like they're just completely incompatible because they use entirely different modalities. If you have a simulator and can plug this all in, you can rule out not only the obvious incompatibilities, but the more subtle incompatibilities: maybe Device A and Device B are both good devices, but their performance isn't great when you put them together just because of the peculiarities about them. It's great to figure this out before you go out and spend money!
- Next, discern which use cases are practically viable on the planned network implementation. So for example, you've gone through all of the steps, you're hoping you can perform certain use cases. This lets you see exactly how feasible that is, and manage expectations for what these early networks are going to be and do and eventually manage expectations for what the full networks are going to do.
- Develop, test, and benchmark protocols for those applications before purchasing hardware. Just like classical networks, there's a whole layered stack of software that helps the hardware perform like a network instead of like a physics experiment. All of that software can be built and developed and tested and fine tuned before you buy the hardware - if you're simulating and/or emulating that hardware. Create the network stack and get it working the way you want it to in emulation first, and then buy your hardware. Then, at the time of your investment in the hardware, you will have something that works, which makes this whole thing a lot less risky!
Once you have all of this foundation laid in the emulation mode, the pilot mode becomes very straightforward. It goes from being this daunting task of creating a quantum network to the very straightforward task of following the plan that you already designed in a very data driven way. So this takes out a lot of the surprises that you would otherwise find in pilot mode, and makes it much more efficient.
What is involved in the second phase of quantum network implementation: Pilot & Trial?
So what we mean when we say a pilot is building a small scale quantum network that's used to test and optimize performance and gain internal familiarity with the technology. This usually looks like quickly building an on-site functional quantum network testbed. This phase gives you an opportunity to test the interoperability between quantum networking devices, optical components, existing infrastructure, firmware, and also software. You can do a lot in simulation and emulation to figure out how things will work together and if they'll work together, but there's no substitute for actually plugging it in and ensuring it works.
If there's any real misalignment, it will come up in pilot mode and allow you to see exactly what that is, and hopefully overcome any snags, mismatches, or bugs before you do a full-scale deployment rollout.
Another benefit of this step is building internal expertise with the technology. There truly is no substitute for having your staff actually touch this stuff and overcome the challenges that come up along the way as they get it working.
A pilot, the tangible result of this phase, is a key component of demonstrating a concrete proof-of-concept and enabling resource allocation for full-scale deployment. It's difficult to acquire a large enough budget to make the quantum network of your dreams immediately, so a lot of organizations consider compromising and making a smaller one that is still kind of what they want for the deployment mode. This is not the best way of going about it, it works a lot better if you scale it down a lot and create a quite small pilot. This way you can overcome all these potential issues, and then use this pilot to show people what is exciting, what is practical, and leverage that to allocate resources for your organization to be able to make the bigger deployment happen - and for it to be as ambitious as it can be.
What happens in the final phase of implementation: Full-Scale Deployment?
This is the final stage: where we scale quantum networks and integrate end-to-end applications. This involves deploying and scaling operational entanglement-based quantum networks. Once you've done that, you can support transformative applications in secure communications, clustered quantum computing, and distributed quantum sensing.
If you maintain a multipurpose hardware agnostic approach throughout this entire process, this final deployment will be very, very future proof. You’ll avoid being locked into specific vendors, specific modalities, or specific use cases that may not withstand the test of time and may not be optimal in the end. Your quantum network can be continually evolving with more sophisticated hardware and protocols, and more sophisticated applications, as they arise throughout the industrial and academic community.
This article is taken from a webinar with Aliro Quantum Product Manager Cara Alexander that originally aired on Bright Talks. You can find that webinar here.
Aliro Quantum, The Quantum Networking Company™, is leading the charge in quantum network development by offering the foundational technologies needed for organizations around the world to build scalable and powerful distributed quantum systems. Aliro also works with industry and academic partners through the Quantum Economic Development Consortium (QED-C), the NSF Center for Quantum Networks (CQN), and the NSF Quantum Leap Challenge Institute Hybrid Quantum Architectures and Networks (HQAN). Additionally, Aliro is involved in several quantum networking standards groups at IEEE and QED-C.
