Slashdot Mirror
MIT's New 5-Atom Quantum Computer Could Make Today's Encryption Obsolete (pcworld.com)
An anonymous reader writes: In traditional computing, numbers are represented by either 0s or 1s, but quantum computing relies on atomic-scale units, or "quibits," that can be simultaneously 0 and 1 -- a state known as a superposition that's far more efficient. It typically takes about 12 qubits to factor the number 15, but researchers at MIT and the University of Innsbruck in Austria have found a way to pare that down to five qubits, each represented by a single atom, they said this week. Using laser pulses to keep the quantum system stable by holding the atoms in an ion trap, the new system promises scalability as well, as more atoms and lasers can be added to build a bigger and faster quantum computer able to factor much larger numbers. That, in turn, presents new risks for factorization-based methods such as RSA, used for protecting credit cards, state secrets and other confidential data. "If you are a nation state, you probably don't want to publicly store your secrets using encryption that relies on factoring as a hard-to-invert problem," said Chuang. "Because when these quantum computers start coming out, [adversaries will] be able to go back and unencrypt all those old secrets."
For an actual summary of this research see http://www.scottaaronson.com/blog/?p=2673 by Scott Aaronson who is a quantum computing expert. The key thing here is that they factored 15 with high probability without having to sort of cheat by making a circuit that was more likely to work if one suspected that 15 had factorization resembling 3*5. As usual, this is getting completely overblown by the popular press. It is an important step towards actually making quantum computers that can factor big numbers, but it is nowhere near anything that would make RSA or other factoring based crypto obsolete.
If you actually read the scientific article (which is available as a preprint unter [1]), what the authors discuss is how to significantly improve Shor's algorithm, the quantum algorithm for factorizing prime numbers. They show that the number of qubits needed to perform Shor's algorithm is actually quite a bit lower than what previous versions of the algorithm required - and they claim that their version is much more scalable than previously known versions.
They demonstrate their algorithm by factorizing the number 15 using trapped ions. That elementary qubit operations can be performed with trapped ions has already been demonstrated [2], that part is nothing new. Factoring the number 15 with Shor's algorithm is has also been done before. But since their algorithm doesn't need nearly as many qubits as the previous formulation of Shor's algorithm, specifically they only need to have a single ancillary qubit in addition to the qubits required to represent the number to be factorized (in contrast to 3n ancillary qubits), and given the fact that the quantum Fourier transform operation that was previously required to be performed on the ancillary qubits is difficult to pull of in practice while keeping quantum coherence, they argue that their algorithm will be much easier to implement in real quantum systems.
So their research is actually a big step forward when it comes to a potential actual practical realization of Shor's algorithm, and what they did is still very impressive (even the experimental part of their work), but their work doesn't address the problem of actually scaling up the number of qubits: 5 bits have been done before, and while their work means that less qubits are needed, it's not like even a (512+1+error correction) qubit computer with quantum coherences is around the corner (note that to break 512 bit RSA you don't need a quantum computer). Furthermore, there's a huge debate in the community as to what the best design for a scalable qubit architecture is: the authors of this paper seem to follow the school that wants to use ion traps, but there are also other approaches to implementing qubits: superconducting qubits (in various variants), spin qubits (including nuclear spins), semiconducting qubits, adiabatic quantum computation, and a couple more. A lot of people in the community are working on all of these different approaches, and it is not clear to me which of these will be the most effective way to implement a quantum computer in the end. And scaling this up beyond 100 qubits with full quantum coherence and quantum control of qubit operations (from all reports e.g. the D-Wave machine "only" does quantum annealing with ~500 qubits, and doesn't implement a universal quantum computer) is something that's still quite a bit away. How long? I don't think anybody can really predict. Could be 5 years, could be 10, could be 50.
To reiterate: the paper is a breakthrough, because (if we leave out error correction for the moment, which increases the number of qubits required) to factor a 1024 bit RSA key, one would previously have needed 1024 + 3 * 1024 qubits and a very difficult to pull off quantum operation (quantum Fourier transform) on 3 * 1024 qubits simultaneously. This paper reduces that to 1024 + 1 qubits, where the KQFT operation only has to be applied to the 1 additional qubit. We still don't know how to actually manufacture a quantum computer that maintains coherence well enough with that many qubits, so there's no need to start panicking when it comes to this, but these kind of improvements do show that research towards asymmetric cryptography that is safe against quantum computing is required - and that we should really start implementing these kinds of algorithms NOW, so that when somebody actually has breakthrough in this regard, we have the technology in place to switch at that point. A good starting point for people that are interested is the pqcrypto.org site [3] and the excellent talk by Dan Bernstein and Tanja Lange at 32c3. [4]
[1] http://arxiv.org/abs/1507.08852
[2] https://en.wikipedia.org/wiki/Trapped_ion_quantum_computer
[3] http://pqcrypto.org/
[4] https://www.youtube.com/watch?v=6XeBvdm8vao
While you could be right that the necessary technology still won't be available in 40 years, the quantum world is fundamentally different from the analog world. In the analog world, noise and other errors determine an absolute limit as to how much precision you can achieve. In the quantum world, there is the miracle of quantum error correction that can compensate for errors. It is quite amazing mathematically that linear transformations performed by quantum gates can correct errors, but the mathematics works (I have worked through it myself, it's not terribly hard, requiring only linear algebra) and small error-correcting qubit circuits have been demonstrated.
Most important is the threshold theorem that says if we can reduce the noise in a qubit below about 1 part in 10^5 (IIRC), error correction can allow a quantum computer to grow to an unlimited number of qubits. That's when the revolution will start.
Quantum computing is dependent on exactly one dubious assumption: That there is no [hard] limit to the complexity of a physical interaction.
If we can have unlimited complexity, then we can have quantum circuits which are as good as [credibly] advertised; if we can not, then, at best, all we get out of it is a means to optimize a few computations.
That key has eluded researchers for a few decades now. It looks very much like there is an upper limit on the number of qbits that can be entangled in practice if computations are to be performed and as if that upper limit is somewhere around 100. With that, not even very old and outdated RSA-768 is threatened.
That is why these stories are so utterly demented. They are akin to claiming the invention of the logic gate will make 2048-bit computers possible that run at 1000GHz. As we now see in practice, 64 bit at 5GHz is pretty much the viable limit for low-cost and it does not go much further with extreme hardware. In reality, things do not scale after a certain limit and for quantum computing, that limit will be very low.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.