Quantum computing is one of those topics that always seem to exist just out of our reach. Well over a decade ago we were promised that a real quantum computer will appear just in a few years, and for most people, this feeling remains the same even now. But is this assumption correct? Do we really need to wait a few years more until the fastest, biggest, and most reliable quantum computers will finally find their way into every industry? Or is it something we should already be considering today?
Understanding (or at least trying to) quantum computing
Now, I will not attempt to explain how quantum computing works in this post: in fact, we’ve already tried that in an episode of the KuppingerCole podcast on this subject. Also, I won’t even pretend I can understand much of quantum physics and its underlying mathematics despite having a degree in it. But one crucial thing that must be reiterated is that quantum computing is based on principles fundamentally different from classical computers and thus is not meant to be a replacement for existing computer algorithms. When reliably working and sufficiently powerful quantum hardware becomes available, it will not magically make existing programs or computational methods run faster.
In reality, the most interesting potential applications for quantum computing are problems that are practically impossible to solve with traditional methods in any feasible amount of time. Various problems in the field of chemistry and nanotechnology are among popular goals; the so-called quantum annealing can be used for solving various optimization problems such as the traveling salesman problem. Finally, the notorious integer factorization problem, if solved with quantum algorithms, can undermine the reliability of public key encryption. This hypothetical advantage over classical computers is somewhat awkwardly referred to as “quantum supremacy”.
Quantum Supremacy – does it even mean anything?
Well, at least it is no longer hypothetical: back in 2019, Google has claimed to have achieved quantum supremacy by solving a problem on a quantum computer in 200 seconds, which would take a classical supercomputer 10,000 years to complete. In December 2020, Chinese scientists have performed boson sampling (I wish I could explain what it even means!) in 20 seconds, which allegedly would take 600 million years of traditional computation. So, does it mean that powerful and useful quantum computers are already a reality?
Unfortunately, it is not that easy. First, demonstrating quantum supremacy does not require solving a practically useful problem. Also, real physical quantum computers are very sensitive to imperfections in the underlying hardware and the influence of the environment. Without introducing special quantum error correction methods, large-scale quantum computing is still practically impossible (think of working with a classical computer that crashes every second). Combined with the enormous difficulties of scaling quantum computers to thousands or millions of qubits needed to implement useful algorithms, it may look like the majority of organizations shouldn’t even worry about practical applications for at least another five years…
Should I start with quantum computing today?
And again, the answer is not that easy. Even if your business isn’t planning to benefit from quantum computing yet, there are some aspects of it that you have to consider anyway. The most obvious one is crypto-agility. Even though the currently used encryption algorithms have not been compromised yet, the chance that it can happen very shortly is high enough. Standards bodies are already working on evaluating and standardizing quantum-resistant cryptographic algorithms, NIST is planning to release their first standard next year already. Vendors working in this area are already implementing the necessary changes in their encryption platforms to make the migration to post-quantum cryptography easier. If your organization keeps encrypted data in the cloud, you should at least be aware of these developments.
But what if you want to start with quantum computing properly? Sure, the current generation of hardware isn’t yet suitable for tasks like cracking public encryption, but surely certain “less glamourous” problems like improving the efficiency of fertilizers and car batteries or modeling financial markets can already be tackled. Where do you get a quantum computer? How do you learn to operate it? Are they available as-a-service?
Yes, they are! In fact, quantum computing as a managed service is perhaps the only sensible option for the next few years. The actual hardware is quite complicated and delicate, requiring a team of experts to operate. Also, a substantial part of today’s quantum applications is running simulations on traditional hardware, which is only feasible at the cloud scale. It is no wonder that all major cloud service providers currently offer at least some kind of a managed environment for researching and experimenting with cloud computing. Google and IBM are perhaps the leaders in developing their own technologies, while Microsoft and AWS have open ecosystems with solutions from multiple 3rd party vendors. Smaller companies like the Canadian D-Wave might offer their own “quantum clouds” as well.
Choosing the right partner
The biggest question is thus not “should I start with quantum computing today” but rather “whom should I choose as the strategic partner”. I’m afraid nobody can currently give a definitive answer for this today as every aspiring vendor has their own advantages. Google, for example, was the first one to achieve quantum supremacy. IBM offers perhaps the most comprehensive long-term roadmap for a full stack of QC services. Microsoft focuses on an open ecosystem with universal high-level tools that can work on hybrid computing architectures. Amazon is attracting developers with a free tier for quantum simulation.
In the end, the choice is entirely yours. One thing is certain though: quantum computing is no longer science fiction. Even today, many companies can already reap tangible benefits by becoming early adopters, hiring the right talent to develop new algorithms and applications, and partnering with the right providers to get access to the actual quantum hardware. Perhaps, with more mature hardware and diverse application scenarios, we’ll end up with a “multi-quantum” reality just like we have with the cloud in general. Will this lead to an emergence of a universal framework for quantum computing similar to Kubernetes? Only time will tell.