This may be the biggest thing in computing since microprocessors. However, quantum computing has no security guarantees. Among the challenges is the development of cryptographic algorithms that can withstand quantum computing. Both governments and industry are pursuing this goal. From an industry perspective, Tom Temin and Federal Drive spoke with Scott Crowder, VP of Quantum Adoption at IBM.
Tom Taemin
Let’s start with a simple definition of quantum computing. This word is on everyone’s lips, but I’m not sure everyone knows exactly what it is.
Scott Crowder
of course. Quantum Information Science, or Quantum Computing, uses the principles of quantum mechanics to do computation in a very unconventional way. And that informatics is exciting. Because it scales differently than how we’ve been doing calculations since we started counting chickens on our fingers. In quantum computing, every additional qubit doubles the state space that can be explored. It goes up really, really quickly. And when you get a quantum computer with around 160 qubits, you get a state space larger than the number of atoms on Earth. At qubit sizes up to 300, the state space is larger than the number of atoms in the known universe. That doesn’t mean you can compute anything. It’s not a powerful computer at all. But I showed the blackboard that I could do certain kinds of calculations that a classical computer could never do.
Tom Taemin
And for those who understand classical binary operations, that no longer applies, right?
Scott Crowder
That is correct. So this is a different approach than the way binary or classical arithmetic works. One of the ideas people like from a popular science standpoint is that if computers get big enough, they can solve any problem. And that is fundamentally not true. And that’s good. Because if that were true, cybersecurity wouldn’t exist. To protect the data, we use problems that are difficult to compute on classical computers. And what is quantum secure cryptography or post-quantum cryptography is something both quantum and classical computers are bad at. So the challenge is that in the future we believe quantum computers will be able to successfully compute the kinds of problems we currently use for cryptography. And quantum secure cryptography or post-quantum cryptography is a math-hard problem for both quantum and classical computers.
Tom Taemin
So the problem or challenge is to either secure current computing algorithms from quantum or secure current computing with quantum-tolerant algorithms. And later protect the quantum itself from itself.
Scott Crowder
yeah, exactly. What happened is that in the mid-90s, a guy named Peter Shore wrote on the blackboard to prove that a future quantum computer of sufficient size and quality could perform discrete logarithm factorization. This is the mathematics we use for public key cryptography today, where basically all of society’s e-commerce works. So I want to send you something secret. But I’ve wrapped it in public-key cryptography so no one else can steal it. Or factorization, etc. A man named Peter Shore proved on the blackboard that the quantum computer of the future could solve a problem that would take billions of years for a regular old computer to take days for a quantum computer.And that’s the challenge. The challenge is that we need to find new cryptosystems that even quantum computers can’t crack. But at the same time make sure that no classic computer can beat it. Because there are many classic computers now.
Tom Taemin
of course. now, [National Institute of Standards and Technology] The Department of Commerce has released guidelines for quantum-safe algorithms. Is that basic or is that what the industry is working on now? Because I think the industry supports their development.
Scott Crowder
Yes, exactly. IBM Research worked very closely with NIST. In fact, three of the four that NIST selected for standardization came from IBM Research and its collaborators. Fortunately, the type of mathematics that seems to fall into both classical and quantum computer categories turns out to be problematic to decipher. And now I want to make sure everything related to that math is solid. Because vulnerabilities can be introduced not only from the algorithm itself, but from the way the algorithm is implemented. The good news is that algorithms exist. Indeed, very smart people have come up with algorithms that are cryptographically secure for quantum computers. But now I just have to make sure I put the standard in place and put a wrapper around it. This will set you up and ready to implement.
Tom Taemin
I’m talking to Scott Crowder, VP of Quantum Adoption at IBM. And despite the occasional claim, how can we really know that quantum computers as imagined do not actually exist? Do humans understand that, or how do we test these with current technology?
Scott Crowder
yes. So I think this is one of the clearest ways you can test theories that are on the quantum side. And the classic side is just as important. In fact, it could be argued that it’s even more important to make sure there are no vulnerabilities that classic computers can discover. Because it is today’s threat. For those of us who have been hacking this for his 10+ years, due to some methods being proposed for standards from NIST. What NIST has done in the past was probably about 7 years ago, so why did they announce that these are the algorithms they want to standardize?
Tom Taemin
Also, can quantum effects be simulated on a classical computer just to run an experiment?
Scott Crowder
They can solve small problems. If you have problems with qubit sizes less than 50 qubits. Yes, it can be simulated on the world’s largest supercomputer. The challenge is that when you go beyond it, you can’t go beyond it. To be clear, today’s quantum systems are neither large enough nor have sufficiently low error rates to be substantially better than classical computers. And I wouldn’t say they’re cryptographically relevant, but we’re progressing very quickly. increase. And we need to prepare for the day when they become crypto-related.
Tom Taemin
yes. What are the challenges to reach the quantum? Is it a manufacturing problem? That means we need processors that work with qubits like Noah’s Ark.
Scott Crowder
Ripens really fast. Come to think of it, I put out a roadmap for the end of 2019, so it was about three years ago. At that point, we were in a 27-qubit system. Now he is 433. This year he will exceed 1,000. It’s one of those practical axes that makes it better and better than the classic. The second most important is the error rate and the algorithms that quantum computers use to mitigate errors. And that’s another axis we need to keep improving. Currently, the fidelity of operations is around 99.9% and the error rate is 0.1%. Once you get a little better than that, you can start error mitigation techniques that make it practical. If you look at our roadmap, we’ll reach about 5,000 by the middle of this decade and continue to drive error rates down to sub-three nines or better than three-nines fidelity. By the middle of this decade, we may have reached a very interesting point where quantum computers will be practical.
Tom Taemin
Indeed, if you have an error of .1, or .01%, or .001% on a scale of thousands of qubits, that is an error of millions. So you have to get these pretty low and then have a very strong way to fix it.
Scott Crowder
That’s right. That is correct. That is the big problem in the industry today. There are always hurdles to overcome. What’s exciting is how much progress we’ve made, staying on the same slope as we are now and starting to intercept in a few years, if not a little sooner.
Tom Taemin
In practice, therefore, manufacturing, programming, error rate reduction, and cybersecurity must go in parallel. In the next decade, for it to become a practical reality.
Scott Crowder
Yes, I would argue that cybersecurity needs to go further than that. The reason we need to do it before that is because we are concerned about collecting data now and decrypting it later. And the reality of the industry is that many industries have platforms that need to survive for the long term. You can think of many similarities in the telephone company network, car manufacturer, or federal space. There are platforms that need to be in the field for 5, 10, 20, 40 years. In the future, when quantum computers are powerful enough, they will hit the field, so we need to start thinking about how to make them quantum-safe now. And obviously for highly sensitive data. The fear is that people can collect data now and encrypt it later. Therefore, this cybersecurity element must lead to keep up with the quantum computing roadmap.