With quantum computers already available for commercial use, albeit not in great quantity, I believe it is time for companies to start considering how to incorporate them into their arsenal of IT services. Time is running out for security teams to get their defenses in order ahead of the first quantum computer attack.
Right now, it’s not really possible to just buy the latest quantum computer model and take it for a spin. It takes some real architectural brain power to actually make such a computational beast fit into an existing structure.
What kind of architect would be needed for such a task, and what would the design look like?
Quantum, Turbo, and the cat
When the first few turbo cars came about, they were marketed by sticking the word "Turbo" on the back. The word became synonymous with "power," so pretty soon, you could find all kinds of household appliances and other gadgets emblazoned with the word "turbo." A similar thing seems to have happened with the word "quantum," as I find it in all kinds of odd contexts like the design of houses or models of working.
"Schrödinger’s cat" is another concept that some people associate with quantum technology. Schrödinger was corresponding with Einstein and used the cat in a thought experiment intended to disprove (in theory) the idea that superposition could only be revealed once observed (before that it could be anything). This idea about how quantum mechanics behaved came from Niels Bohr and Werner Heisenberg and is often referred to as the "Copenhagen interpretation."
NB: A recent article from Jussi Lindgren and Jukka Liukkonen "concluded that the correlation between a location and momentum, i.e., their relationship, is fixed. In other words, reality is an object that does not depend on the person measuring it." To me, this is vastly cool because it continues to fuel a discussion that started 100 years ago!
[ You might also enjoy: Sysadmin fails: When service dependencies go wrong ]
In-house quantum computer
This is not to say that each company has to have their own quantum computer; that’s a difficult order, partly because they are quite expensive and rather delicate to maintain, but also because of the blistering pace of progress. The predominant risk of a quantum computer being outdated even before it is properly run in is a bit daunting and should spur the plans to source such a service.
Also, consider that today's quantum computers don’t really have an equivalent to Windows or Linux—a generic operating system that works on many different quantum computer hardwares. Of course, there are projects that have this aim, but I believe we will need to go through a few generations before some sort of standard emerges.
Interestingly, the in-house solution for a quantum computer would require much of what we saw in the early days of mainframe computers. We’re talking lavish location, comprehensive cooling, excessive electrical consumption, draconian cleanliness, expert access only, and of course, no external interference whatsoever. If this sounds like your company, then, by all means, go ahead with an in-house quantum computer. Otherwise, you’ll probably want to consider QCaaS—Quantum Computing as a Service.
QCaaS, speed, and trust
After determining your trusted service provider, you’re now challenged with how to integrate this service-beast into your current IT landscape. From an architectural point of view, it is a good idea to collaborate with the security team on this. Connecting to a quantum computer also means that it can connect back to you, but with a significant difference in processing power. Exercise caution since you don’t know who else is using the same quantum computer at your provider. Security is best served as a starter, not as a dessert.
I have previously spoken about how quantum computers can use standard network components to communicate and can very well exploit their weaknesses. So it is time to gear up, or else you will have to trust that your provider has taken ALL necessary security measures and that ALL their customers can also be trusted. As the adage goes, "trust, but verify."
What kind of architect?
When it comes to IT architects, there are several areas of expertise within this discipline, and I thought I’d mention just a few in order to set the landscape.
This is where the bold plans should start. The bigger picture should, of course, be driven by business requirements and fuelled by the ever-changing market. Your enterprise architect should present plans for areas where quantum computing would bring a competitive edge to the business. For instance, the healthcare industry is all over quantum computers because of their ability to do what classic computers can not.
The application architect will have a more in-depth challenge since the details need to be uncovered. How will the quantum computer interact with the legacy systems? Where in the flow will the quantum computer sit, and how will the flow react to binary questions receiving "non-binary" answers? Perhaps that is an incorrect way of visualizing the problem—after all, it is not the answer that is non-binary but rather the process of obtaining the answers. Either way, I think this adds some really interesting flavor to the process, don’t you?
Next, there is the quantum application and how to develop it. Perhaps you don’t have to learn the mathematics and programming of quantum computers to get a competitive edge, but I bet that those who do will have a definite advantage. Then there is the argument that your advantage is all about the data you own and the questions you ask. So perhaps it will be enough to use the applications available from your QCaaS provider. The application architect will be the one to provide the answers here.
Is the competitive edge in the actual programming, or is it in asking the right questions and "just" providing the data for the quantum computer to work on? What I have been reading has made me consider the power of asking the right question, especially if you consider a quantum computer with AI capabilities. Currently, computers only do what we tell them to do. AI is seriously hampered by the lack of processing power in classic computers, so it is only reasonable to assume that a quantum computer will not only unlock that obstacle but also open an endless space for humankind to ask the “right" questions. In case we have not considered ALL the possible consequences of our question, we may find that it is interpreted in a way we failed to anticipate. Can an architect design enough safety measures to protect us from our own inability to foresee the future?
For the data architect, the challenge becomes even more granular since the two environments—quantum and classic—will relate to data in very different ways. Where the classic computer has the approach of a Sith Lord and deals in absolutes, the quantum computer is more of a Jedi that can feel the flow and use the force to find the answer. Again, how to have the classic data interact with the quantum data is a dilemma that makes my mind spin. The same data will be processed completely differently in these two realms, which means the output needs to be managed very carefully.
Since quantum computers can process huge amounts of information and, at the same time, try out multiple solutions, the question then becomes how to provide all this data to the quantum computer. Even with fast disks, it is the equivalent of having a supercomputer accessing a 1.44 floppy disk. I would expect some major rethinking of data store and access times. And this is before we consider data security and integrity. Most likely, the necessary tech just hasn’t yet been invented.
The cloud architect faces the dilemma of how to get information to and from the quantum computer. Everything costs in the cloud and that includes the transport of data. Depending on what type of operations the quantum computer performs, it might need huge amounts of data in order to explore all possible solutions. To feed all this data across the cloud might become expensive, but perhaps the business benefit will significantly outweigh the costs.
On the other hand, I’d expect that cloud would take on a new dimension if we employed quantum computers that are orbiting the earth. Successful experiments have already been undertaken in space; however, even if there is plenty of vacuum and an abundance of coldness, there is still the challenge of interference—but no doubt we will soon see that sorted. So I think, being a cloud architect, you have the potential to develop your role in a very interesting direction!
The challenge of the security architect is not to be underestimated. If you invite a shark for dinner, don’t be surprised if it consumes a lot more than what’s on the table. What I am alluding to here is the fact that, since most companies will go for QCaaS, this might open a two-way street between the provider and the customers. The encryption and security protocols we have today have brought forth some scaremongers claiming that it will be a walk in the park for quantum computers to break any and all security available in a matter of seconds. Yet others claim that we do have sufficient security already and that there is no cause for concern. I guess only time will tell.
Only a few companies have taken real actions when it comes to data protection with specific regard to quantum computers. There are several initiatives working on improving the encryption algorithms to ensure that even a quantum computer of rank could not break. You can search for "post-quantum cryptography," and you’ll find plenty of opinions and facts on the internet—just don’t get the two mixed up. Having studied some basic security, I learned that “all it takes is time” and that the quantum computers of the near future will be so powerful that they can literally bend time. I wonder what will happen then?
The network architect is the ninja of them all. "Think quick, act quicker" is the neighborhood tagline. Since quantum computers can use the fiber optics we have today for communication, there is also a need to review the configuration items in the network as well as the monitoring tools. Looking at TCP/IP network transfer is one thing, but how do you monitor individual photons zooming across the lines? Some new components have to be introduced for all this to work, one of them being “quantum repeaters”—tech that will be unbelievably busy. Having said that, much of the network technology will be standard.
Having quantum computers that compare data or work with "data lakes" can put some serious pressure on the network and, considering the speed at which quantum computers work, the need for some serious data transfer is not just on the horizon but kicking the front door down!
With the capacity to access and plow through huge amounts of data and provide multiple answers in a veritable vortex of processing, the network lines benefit from being as clean as possible—meaning not a lot of other traffic or disturbance. The photons transmitted between quantum computers need to have their information "repeated" after a certain distance. How to manage and feed all the control information to the consoles that the network technicians work with will be a huge challenge. How can you monitor the progress of something that is already done before it started?
What type of Quantum Computer do you need?
Quantum computers do come in different flavors, and I would expect even more to emerge as the technology continues to mature. Depending on what your company needs, you may opt for a specific type of quantum computer. Check out the following variations:
Quantum annealing is best for solving optimization problems—like optimizing the traffic flow for each individual vehicle in an entire city (like Beijing). This is something that could be performed in real time and continuously provide instructions to each individual car. Less congestion, shorter travel times, and less environmental impact would be the immediate outcomes.
Quantum simulations explore specific problems in quantum physics that are beyond the capacity of classical systems—like specific pharmaceutical challenges. For a new drug to be discovered using conventional computer power, there is, on average, a 90% failure rate in the clinical trials. And for it to reach the market, the cost is around $2 billion over a ten year period. An American company called FAR Biotech is now in the process of "identifying drug-like molecules throughout the chemical space of about 1.5 trillion chemical structures (including new chemical entities, known compounds, and repurposing drugs) and via accessing the data of 200+ models of human therapeutic targets." All this will be done in a fraction of the time it takes a classical computer to even reach the first failures. Here, you can really see the competitive edge.
Universal quantum computers are the most powerful and most generally applicable where you can target any massively complex computation and get a quick solution—here is where Artificial Intelligence is looking for a boost. I could imagine that NASA is looking to employ this type of quantum computer to track the approximately 20,000 pieces of larger space debris that’s orbiting the Earth and pose a serious threat to spacecraft. There are also some 500,000 pieces of smaller debris the size of a marble or larger, and there are many millions of pieces of debris that are so small they can’t currently be tracked. But a quantum computer would have the capacity to do so.
Quite reasonable answers
Like I said before, our current computers are just like a Sith Lord—they deal in absolutes in the sense that it’s binary: either 1 or 0. The quantum computer can deal in both. It can actually deal simultaneously with all possible versions in between 1 and 0. The outcome is a probability, which then makes the interaction between a classical and quantum a playground for misunderstandings. Will the classical system be able to pose a question and, in return, accept an answer of what is most likely? I can see how IT architects need to have a much deeper understanding of how these two worlds differ and the conditions necessary for these very different worlds to interact safely and consistently.
Being safe together
Perhaps you are an IT architect and do the boxes-and-lines-design when integrating a quantum computer with the existing IT landscape, and then let the individual teams break down the details. It could work, but I suggest you engage at a more granular level.
A quantum computer is like having a car with a million horsepower that can drive straight through any obstacle, fly and dive at the same time, plus it can continuously change color and shape. Now try to put that on an ordinary road and have it interact with regular cars. Needless to say, this will not work without some very clever design that keeps both parties safe from one another but still able to perform the job of transporting people between different locations.
[ A free guide from Red Hat: 5 steps to automate your business. ]
Many times I have been impressed with IT architects when they are presenting something new in a way that is both elegant and compelling. To design and explain something "futuristic" but still with a clear path of how to get there and why it makes sense is the greatest gift. As sysadmins, we have to turn plans into reality, but quantum computing goes beyond what we know and understand currently. A quantum computer is a different beast from any angle, but the challenge is that it is literally all angles at the same time. Quantum Computing as a Service will most likely be the start of how companies will employ these services, but setting it up in a secure and useful way will be a challenge.
What kind of quantum computer to use is yet another challenging aspect, as well as how to feed it with enough data to come up with the most probable answer. How we ask questions to a quantum computer will determine what kind of answers we get. The only recommendation I can offer is: learn about quantum computers, how they evolve, what they can do, how they communicate, and what the opportunities are as well as the risks. Quantum computers are here now, and, in a few short years, there will be even more companies turning to QCaaS. Will you follow the crowd, or will you lead?