backtop


Print 38 comment(s) - last by Diesel Donkey.. on Aug 23 at 2:45 PM

Scientists use pulses of light to dramatically accelerate quantum computers

Quantum computers are known for being able to solve complex problems in far less time than traditional computers. Cracking encryptions could be one area in which quantum computers excel.

By using pulses of light to dramatically accelerate quantum computers, University of Michigan researchers said that they have made strides in technology that could foil national and personal security threats. It's a leap, they say, that could lead to tougher protections of information and quicker deciphering of hackers' encryption codes.

Optically driven quantum computers can crack highly encrypted codes in seconds. The fastest of today's desktop computers would require 20 years.

The researchers used short, coherent pulses of light to create light-matter interactions in quantum dots — particles so small that the addition or deletion of electrons changes their properties. They found they could control the frequency and phase shifts in the optical network, which is crucial in powering an optically driven quantum computer, noted researcher Duncan Steel.

"Quantum computers are capable of massive parallel computations," Steel said. "That's why these machines are so fast."

"We're particularly excited about our findings because they show that we can achieve these results by using quantum dots and readily available, relatively inexpensive optical telecommunications technology to drive quantum computers," Steel added. "Quantum dots replace transistors in these computers, and our results show that it only takes a few billionths of a watt to drive it."

The paper on the results of this research, "Coherent Optical Spectroscopy of a Strongly Driven Quantum Dot," appears in the August 17 issue of Science.



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

pretty cool...
By Moishe on 8/20/2007 9:43:17 AM , Rating: 3
Cool, and scary. Between the time when the government gets this and when it becomes cheap enough for regular users/businesses to get it there will be a huge potential for government abuse of privacy. People will really not be able to expect any privacy online at all.

not that I'm particularly paranoid or have anything to hide... but it's still almost never a good thing to have something so powerful in the hands of a few unaccountable people.




RE: pretty cool...
By omnicronx on 8/20/2007 9:57:46 AM , Rating: 1
Did you really think the little lock showing 'page encrypted' ever stopped anyone before? If it were legal the government would already be all over us, and really is there any online privacy to begin with?

my rule of thumb, if you are going to do something illegal online, watch your back! because someone is always watching!


RE: pretty cool...
By FITCamaro on 8/20/2007 10:56:50 AM , Rating: 6
A better rule of thumb would be to not do anything illegal.


RE: pretty cool...
By Kuroyama on 8/20/2007 1:37:24 PM , Rating: 3
There are plenty of authoritarian countries in which you don't have to do anything illegal to be disappeared. And many things are illegal only to preserve existing authoritarian regimes, so sometimes violating such laws is a good thing. In such countries encryption provides one of the few means of safely communicating things that the government doesn't want people to discuss. Once it reaches the market quantum computing is not going to be the sole domain of the US and other governments of free countries. Hopefully, unbreakable quantum encryption will not be far behind.


RE: pretty cool...
By Omega215D on 8/20/2007 5:20:06 PM , Rating: 3
Define illegal.

Sometimes people in power can create laws that don't make any sense.


RE: pretty cool...
By rdeegvainl on 8/21/2007 2:32:59 AM , Rating: 2
That's right, and you South Carolinians, only missionary for you. If government gets word that you partake otherwise it's over!!!


RE: pretty cool...
By Master Kenobi (blog) on 8/20/2007 9:59:06 AM , Rating: 3
I will put down good money and say that the NSA is already going over this and figuring out application for it. The CIA likely isn't far behind. This has been persued by both agencies for a long time, and they are funneling a nice pool of money into it's development.


RE: pretty cool...
By KristopherKubicki (blog) on 8/20/2007 1:04:35 PM , Rating: 2
I've said for a long time -- the US government almost certainly has the quantum computer, and it's probably guarded more secretly than The Bomb was.


RE: pretty cool...
By DokGonzo on 8/20/2007 4:05:05 PM , Rating: 2
Yes, and it's name is SKYNET.


RE: pretty cool...
By Zurtex on 8/20/2007 4:36:53 PM , Rating: 2
Skynet is actually the communications system that the British Military uses: http://en.wikipedia.org/wiki/Skynet_%28satellites%...

:P :P


RE: pretty cool...
By lompocus on 8/20/2007 10:49:22 PM , Rating: 1
They didn't have quantum computers. I know this from stuff I can't tell you lest you die xD. Proof: hacks from others into government files.

BUT! I CAN tell you the regular non conspiracy (lol @ you stupid conspiracists like kristopher) has had quantum computers for a looooong long time. Its just now that they're better than modern computers. Take q-wave: around 10 years late in the quantum race, but they are a company that has shattered, burnt, and wiped their ass with moores law; 24 qubits (?) to 4000 qubits (?).

So kris...now do you realize that not everything is a conspiracy?

I CAN tell you we've had invisibility technology since somewhere between the mid 80s and the mid 90s, though :P. The information has been widely available for a loooooooooooong time. you just need to google the right keywords to prove my point.


RE: pretty cool...
By Diesel Donkey on 8/23/2007 2:45:25 PM , Rating: 2
If that is the case, then the US government certainly is doing a good job of covering its tracks by shelling out tons of money to physics labs like the one in which I work for research that merely touches the very edge of true quantum computing. Of course there are many ways to define quantum computing, and some of them are currently feasible although decidedly a victory for semantics as opposed to actual advancement in computing capability.


RE: pretty cool...
By Shining Arcanine on 8/20/2007 10:46:28 AM , Rating: 3
Do you honestly think that the federal government has the time to abuse privacy when it is using all of the computing hours that it has cracking encryption used by terrorists?

If any government has the time to abuse privacy, it will be the state government, but it is strange that privacy is the first thing that comes to your mind when you read about new computers that can crack encryption. When I read the article, I thought that this would enable the federal government to crack the encryption protects files on the captured terrorists' hard drives.


RE: pretty cool...
By Moishe on 8/20/2007 11:07:25 AM , Rating: 2
My point is this. If they have to use all of their computing time to crack encryption for terrorism (or theoretically "good" uses) then what happens when encryption cracking is no longer time consuming? At that point, all knowledge is practically free, whether it's terrorist or your email or back account.

Another way to put it is that I don't trust the government to do the right thing. If there is any illusion of privacy now I think it is simply because there is not enough time to invade privacy AND crack enemy communications. They choose the obviously higher priority targets. If all targets could be cracked quickly, I guarantee that there would be more snooping on what we now deem to be unimportant targets.


RE: pretty cool...
By Master Kenobi (blog) on 8/20/2007 11:39:52 AM , Rating: 3
You need to go work on an NSA or CIA contract sometime that involves this sort of work. I've done a few (Currently working on one for DHS) and let me just be perfectly blunt and say that you have nothing to worry about from the Government pulling up all your information and finding out stuff about you. I would be more concerned with the grocery store tracking my purchases using those "Bonus Cards" than any federal agency. Until you get to see the classified spaces where they track and store and arrange all the collected data, (Or in most cases, the absolute clown circus and joke it is) I suppose you can't see just how far fetched these conspiracy theories are. You guys give the government way too much credit.


RE: pretty cool...
By Zurtex on 8/20/2007 11:57:23 AM , Rating: 2
I'd be interested to see what the American government is working on and see how far advanced they are.

I know a couple of people (I live in the U.K) who have been on a course at GCHQ and have been tempted to go on the same course myself some time, they are the secret service communication headquarters. They came up the type of encryption we use today on the internet about a decade before it was commerically discovered, they're in to some crazily advanced stuff these days, they've been using quantum communication at least for 10 years.


RE: pretty cool...
By rsmech on 8/20/2007 12:03:20 PM , Rating: 2
quote:
I would be more concerned with the grocery store tracking my purchases using those "Bonus Cards" than any federal agency.


Correct. What the gov't has,IF it has anything is nothing compared to what you could buy from any of these sources.


RE: pretty cool...
By Moishe on 8/20/2007 12:26:09 PM , Rating: 2
ohh yeah, the critical stuff, like what I eat :)
That's all they know.

Now... merge all of the databases and we start having real trouble.


RE: pretty cool...
By Ringold on 8/20/2007 4:48:05 PM , Rating: 2
quote:
merge all of the databases and we start having real trouble.


They can't even merge their damn e-mail databases!

WalMart, on the other hand, knows precisely what specific flavor of poptarts to stock in advance of a hurricane.

Come on. You saw a random chance to take a pot-shot at evil imperial government doing what government has always done in America since *at least* 1861, if not 1776, and that's keep as close an eye on its subjects as it legally can, especially any that may be falling by the wayside.

And what happens in other countries is, as an American, not my concern if the local African despot sells some oil rights to buy a quantum computer to see what pr0n his school boys are looking at on their OLPC's.


RE: pretty cool...
By Ringold on 8/20/2007 5:06:07 PM , Rating: 2
Not that you're entirely wrong, I'll add. I just don't see anything more sinister about it; rather, such capability would merely be keeping up with the times. In theory unlawfully gained knowledge on their part couldn't be used against you anyway; at best it'd maybe be used to identify those who need to be looked at more closely so as it get admissable evidence?


RE: pretty cool...
By Zurtex on 8/20/2007 11:37:35 AM , Rating: 2
lol, is this a joke post?

Anyway there isn't really any personal encryption out there that governments can't crack. Unless people manually write up their own RSA encryption codes using primes of about 1024 bits or use some legally dubious software to do similar. Not submitting how you encrypt something to the government is illegal.

Quantum computers are quite cool as they are a super set of normal computers (i.e they can do everything a normal computer can and more). There's a lot of really good quantum algorithms starting to come out as theoretically computer scientists start thinking in quantum terms, i.e you no longer have 1 and 0, but rather have the probability of having a 1 or the probability of having a 0.


RE: pretty cool...
By omnicronx on 8/20/2007 11:44:56 AM , Rating: 2
quote:
i.e you no longer have 1 and 0, but rather have the probability of having a 1 or the probability of having a 0.
Is this what quantum computing really is? my girlfriend tried to explain it to me but to no avail. (shes a physicists heh)


RE: pretty cool...
By Zurtex on 8/20/2007 11:52:52 AM , Rating: 2
quote:
Is this what quantum computing really is? my girlfriend tried to explain it to me but to no avail. (shes a physicists heh)


Haha, yep, that's pretty much how it works. Though the pragmatics of writing software and building hardware isn't exactly simple.

2 of my house mates were physicists last year and 1 of them was studying quantum computing, very difficult subject but fascinating. Being a mathematician I got to grips with the hard bits quite quickly, but there's still some conceptual jumps you need to make from normal computer to understand what someone is about when they are talking.


RE: pretty cool...
By TWags on 8/20/2007 7:31:24 PM , Rating: 2
quote:
you no longer have 1 and 0, but rather have the probability of having a 1 or the probability of having a 0


Actually, you still have 1's and 0's defined by the spin of the electron relative to the ion's nucleus. However, you also can have the superposition of 1 and 0 simultaneously. Personally I think "spin" was a fairly poorly coined term for it since it's quite loosely based on angular momentum. In any case, using lasers of specific wavelengths you can excite only the electron only if it is a specific state. The probability it will become a 1 or 0 when it falls back is still there, but within milliseconds (if not less) you've set it to what you wanted.

I also don't see the problem of cryptography. Most of modern cryptography just uses the fact that computers are very slow at things like factoring of large prime numbers. I'm sure once we get quantum computers to have enough qubits to do any sort of big calculations (you need about 50 qubits per bit for redundancy's sake if I'm not mistaken, so we're decades away) that it will have switched and begin to utilize newer methods, whatever they may be, that are as much of a pain to crack with quantum computers as current encryptions are with everything we have.


RE: pretty cool...
By Zurtex on 8/21/2007 8:56:35 AM , Rating: 2
Hmm, yes I agree with you, but there is going to be that period where scientists and the military have fully functioning quantum computers and the commerical and consumer level are still using sequential and parallel processors.

There's lots of interesting things coming out of quantum and elliptic cryptography, I'm sure they'll create something similar. It did take a while for the current RSA cryptography to be made, but then I think a lot more people are focused on making a new one for quantum computers.


RE: pretty cool...
By rsmech on 8/20/2007 11:59:55 AM , Rating: 2
quote:
unaccountable people.


They are not, they are elected or should have the oversight of elected official. If there is a problem with the Government, their is a problem with the voters. We all choose who stays & who goes. We are just happy with the same old story every election. The problem is unlimited terms for congress. If it was so important to put term limits on the president to limit his powers, why not the congress. Some of these people have some deep connections & we are the ones that allow it. The only abuses of power they could have are the ones we let stand next election cycle.


RE: pretty cool...
By Moishe on 8/20/2007 12:29:36 PM , Rating: 2
That's funny... maybe all these years of voting and seeing only a worsening of corruption has made me a bit jaded.

It certainly appears like they can do anything they want (within reason) and get away with it. I'm not seeing any increase in integrity amongst government workers.

Add to that the fact that most of the government is not elected, I'm not feeling any safer.


RE: pretty cool...
By Ringold on 8/20/2007 5:03:26 PM , Rating: 2
quote:
I'm not seeing any increase in integrity amongst government workers.


Goldwater had it right.


RE: pretty cool...
By Treckin on 8/20/2007 9:17:51 PM , Rating: 2
as if the government doesnt have one of these sitting in fort mead... that would be a terribly ignorant thought. As if the government technology at the highest levels and upper echelons of security arn't above what ever college level researchers and tech bloggers can conceive...


By geddarkstorm on 8/20/2007 2:27:13 PM , Rating: 2
It sounds like such a system as this could easily lead to seriously quantum computers, maybe even to consumers. But has anyone made a real quantum computer yet? These articles talk as if they are there, but all this is is theory and proof of concept, not working designs. Still, it's a crucial and much needed step.




By Zurtex on 8/20/2007 2:43:31 PM , Rating: 2
Quantum CPUs have been made. They're just in a very primitive state at the moment. IBM showed a quantum computer factor 15 in to 5*3 in faster than polynomial time :-P a few years back.

It's going to be a little while before they get down to the personal consumer level. Though a specific quantum computer can now be commerically bought:

http://www.dailytech.com/article.aspx?newsid=6102

Though my understanding is this isn't quite the general quantum computer everyone is waiting for.


By Kuroyama on 8/20/2007 3:39:21 PM , Rating: 2
quote:
factor 15 in to 5*3 in faster than polynomial time

Presumably you mean they used Peter Shor's algorithm? This is rather nitpicky, but I can factor 156456419813212657 in constant time by hand, although the constant may be millions of years. And the FFT method of multiplying large numbers is "more efficient" than "standard methods", but only if the number is say 150 digits or more. Even linear programming seems to be best solved by the simplex algorithm, despite the existence for over 20 years of "more efficient" interior point methods (although I hear they may be competitive now). Point being that large scale quantum computers might be meaningless unless they are sufficiently fast; although this article seems to suggest that will be the case once they get out of the lab.


By Zurtex on 8/20/2007 4:46:39 PM , Rating: 2
Yeah, I think it was the Peter Shor's. I've not done too much study in to this, just the odd little bit of reading here and there, I get the general concepts but would be terrified to try and write an algorithm myself as it stands.

Of course talking about exponential, polynomial, logarithmic time etc. is all a bit silly with just 1 example.

Also a bit silly if your comparing 2 time systems and your dealing in "small numbers" for example:

If your polynomial time is something like 1/2(n^2 + n) and your logarithmic time is something like (10^32)*Ln(n) and your only interested for n < 100'000'000


The weak point of any system...
By jskirwin on 8/20/2007 11:37:40 AM , Rating: 2
Is any person using it.

Sure the message will be safe while it travels, but any person receiving it will no doubt leave his/her password on a postit note in his cubicle.




RE: The weak point of any system...
By Zurtex on 8/20/2007 11:48:50 AM , Rating: 2
The type of encryption this is talking about cracking is where you have 2 large prime number (think 200 - 300 digits long) say p and q and you multiply them together to get p*q.

The publicly known key is p*q and the private keys are p and q. If you know p*q, you can lock a document, and if you know p and q, you can open a document.

Normal computers can't work out p and q from p*q in a "fast" amount of time" i.e the larger the number, the MUCH larger the time it takes to work it out. Quantum computers can work these problems out much more quick, larger numbers only take a bit longer to work it out.


forget about conspiracy theories
By bigpow on 8/20/2007 4:48:03 PM , Rating: 2
I'd like to see this being used to help humankind, ie: folding protein, Human Genome, etc




RE: forget about conspiracy theories
By Zurtex on 8/20/2007 6:17:11 PM , Rating: 2
You probabily will as well, in order of things which first get used by new and high end computers, is science, military, commerical, consumer. Though the first 2 are some times interchangeable, computers are almost always designed and created originally to solve some scientific or mathematical problem.


proof of concept?
By Kuroyama on 8/20/2007 1:29:45 PM , Rating: 2
I don't have access to the article, but this sounds like just a proof of concept. Why then all the posts about how "scary" this is; it doesn't sound any "worse" than any other advance in quantum computing.




"If you mod me down, I will become more insightful than you can possibly imagine." -- Slashdot











botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki