Slashdot Mirror
New Research Cracks AES Keys 3-5x Faster
Landing his first accepted submission, qpgmr writes "AES, generally thought to be the gold standard for encryption, is showing weaknesses. From Computerworld: 'Researchers from Microsoft and the [Belgian] Katholieke Universiteit Leuven have discovered a way to break the widely used Advanced Encryption Standard, the encryption algorithm used to secure most all online transactions and wireless communications.'"
The full paper has lots of details. Note that it would still take a few billion years with current computers to actually break anything, but there may be further vunerabilities yet to be discovered.
The Katholieke Universiteit Leuven (KUL) is a Belgian, specifically Flemish, university not Dutch.
If all else fails, immortality can always be assured by spectacular error.
Or it would only take a year with a few billion computers.
No, they just use Keyloggers.
"New Research Cracks AES Keys 3-5x Faster"
(the fine print)
"it would still take a few billion years with current computers to actually break anything.."
linky...
Lorem ipsum dolor sit amet, consectetuer adipiscing elit.
To put that number in perspective, it would take a stack of 4GB hard drives extending past the orbit of Saturn...
you mean our equipment?
it's widely known that the NSA uses all known operating systems for distributing computing tasks.
So every windows computer connected to the Internet will accept NSA task packets and compute them and send them back. It does this seamlessly though so the user never sees anything. They built it into the TCP/IP stack. It just becomes easier with windows and even Linux. (SELinux anyone?)
They're using their grammar skills there.
It is usually not practical to pick "a simpler, more elegant math algorithm" because those are easy -- or at least easier -- to break. As someone mentioned up-thread, and as Bruce Schneier likes to remind us, attacks tend to get better over time -- they never get worse.
Modern cryptosystems are carefully tuned to resist a lot of clever attacks. Probably every stage in every (credibly) proposed encryption scheme has been closely examined by very smart people to understand its behavior and look for weaknesses. Existing systems have very elegant structures that are simple in most respects, but they are complicated in certain ways because consistently simple designs are much easier to exploit.
(As a further complicating factor, using a longer key generally requires using more internal state and more rounds. You might -- or might not -- be able to double the block size, but to move from 128-bit to 256-bit keys, you are very likely to need to increase your cipher from [say] 8 rounds to 12. This means at least a 50% increase in execution time for the same amount of data, and possibly more. If the increased size bumps your S-boxes, state and code out of L1 cache, it will be much worse. If you cannot double the block size, but need to double your internal state size for the larger key, that will add another doubling of execution time.)
An interesting observation. Though it is dampened by the fact that brute forcing encryption is pretty much the poster child of an application that lends itself well to parallel processing.
The NSA called. They deny that any such data center exists.
No. To crack AES-128 the attack still requires work of the order of 2^126.1. A machine capable of cracking a 56-bit DES key in a second might be built for about US$5B, going by the price of the COPACABANA FPGA-based DES cracker (US$10,000 for a machine that can crack 56-bit keys in 6 days). Such a machine would take 140 trillion years to crack AES-128 by brute force, or 38 trillion years to crack AES-128 using the algorithm. If you had 38 trillion of these machines you could conceivably crack an AES-128 password in a year. But to give you some idea of how big 38 trillion is, if each of these 38 trillion machines could be made to fit in a 1U server box, the rack would be just over 1.672e8 km high, just a bit over one astronomical unit. You could build a bridge from the earth to the sun with that. If you spread that many machines out, they'd cover 8,892,000 square kilometers, which is more than the total area of the lower 48 states of the US, and you'd have enough machines left over to pave over just about half of Alaska. If they ran at 100 W each, the project would require 3.3288e16 kWh of energy, or 1.2e23 joules, about a thousand times more than the world's annual energy consumption.
For 256-bit keys the problem is even worse. The algorithm has a complexity of 2^254.4. The energy requirement of that staggering number, assuming a computer able to operate at the von Neumann-Landauer limit of ln(2)kT energy per bit flip, running at a temperature of 2.7 K, would require a staggering 1.24e54 J of energy, about the equivalent of 10 billion supernovas, or about a thousandth of the total mass-energy of the Milky Way Galaxy.
Qu'on me donne six lignes écrites de la main du plus honnête homme, j'y trouverai de quoi le faire pendre.
If you choose to believe some of the articles, it was Microsoft who "broke" this encryption algorithm.
However, if you read the actual research paper the first page explicitly explains the relation between Microsoft and the researchers as "The authors were visiting Microsoft Research Redmond while working on these results."
echo '[q]sa[ln0=aln80~Psnlbx]16isb572CCB9AE9DB03273snlbxq' |dc
What do you mean by "known to be secure"? Do you mean that nobody knows how to break it or that there is a formal proof that no shortcuts for brute force attacks exist?
http://www.moonlight3d.eu/