2

How many years from now will we stop using aes encryption algorithm, as of now aes seems to be virtually immune to brute force attacks and not showing any weaknesses, will quantum or optical computing spell an end to aes? Sorry if any mistakes in question or sentence formation Thank you

closed as primarily opinion-based by Mark, schroeder, RoraΖ, Graham Hill, Iszi Jun 22 '15 at 18:30

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.

  • is this a duplicate of "how can i see the future"? – Skaperen Jun 21 '15 at 8:50
  • No, what is that, "how can i see the future" – Lokanath Jun 21 '15 at 8:52
  • 1
    It was a joke. You are asking for a prediction no one can make with confidence. – Z.T. Jun 21 '15 at 11:10
  • Oh, but one can predict hypothetically, I don't want the accurate answer, – Lokanath Jun 21 '15 at 11:14
4

A quantum computer, if it ever exists, ought to be able to explore a space of N possible keys in time √N. I.e. it could brought down the cost of brute-forcing a 128-bit AES key from the technologically infeasible 2128 down to the much lower but still very substantial 264 -- it will still be quite expensive, especially for a machine that does not exist yet. Note that this contrasts a lot with quantum computers trying to attack a RSA or DSA key: keys for asymmetric algorithms are mathematical objects with a lot of internal structure, and "breaking" them amounts to unraveling that structure; AES keys do not have any such structure, they just are bunches of bits.

Decoherence is the biggest hurdle with quantum computers; to build a QC, you need to put together and entangle "qubits", and make them transition between states without losing their links. To make an analogy: imagine that you are trying to put in the same room a thousand hip-hop dancers. They must all make the same moves in perfect unison. The show stops as soon as any one of them begins to make suggestive gestures purporting to characterize the morality of the ancestry or one of his fellow dancers. If you can avoid that for a full month, then congrats ! You have a working quantum computer. All you then need is to keep it going for a century or so, and you will break AES.


Optical computing does not even reach that level. A "fully optical computer" is classical in nature; for such a computer, AES brute-forcing is still a 2128 effort.

There used to be some hopes about faster computers by using optical gates, based on some preliminary analysis about bistable optical elements that could switch between states about 100 times faster than the corresponding electronic gates. This, however, does not translate to 100 times faster computers:

  • These measures were from more than 20 years ago. In the meantime, electronic gates were much optimized, and the gap has shrunk.
  • In modern computers, the gate switching time is no longer the bottleneck; what limits computing speed is the time needed to convey the information from one gate to the next. The problem is the speed of light. An optical computer does not solve that, since light goes, after all, at the speed of light. At best, a fully optical computer can help a bit in reducing the average path length between two gates because light rays can cross each other freely. Don't expect too much of a boost, though.

Moreover, even if you could make a computer that is 100 times faster than existing technology, then you would gain less than 7 bits in brute-force (because 27 = 128). A 2121 effort is far beyond what can be done with existing and foreseeable technology.


We are still using DES, if only as part of Triple DES, despite the publication of AES as a standard in 2001, and the deprecation of DES. Old technologies never die out; at one point we begin to use new ones, but not as a full replacement. Therefore, one may surmise that we will use AES forever, for (at least) backward compatibility.

Not the answer you're looking for? Browse other questions tagged or ask your own question.