Quantum cryptography goes a step further than electronic cryptography through its employment of a stream of photons, the quantum properties of which determine the key. The fun part is that if an intruder observes or intercepts the transmission, those properties get changed -- an unavoidable principle of quantum mechanics -- meaning the sender and receiver can tell if anyone is eavesdropping. Perhaps more important, the key can't be copied or faked (see Optical Science Gets Spookier and Quantum Cipher Sent by Fiber).
It's a potential breaththrough, though working with photons has never been easy, and, as the optical networking bubble has shown, it can be an expensive way to build technology.
MagiQ's Navajo system, a box made to fit in a standard telecom rack, was unveiled in February and began beta trials in March (see MagiQ Demos Quantum Cryptography).
MagiQ says Navajo performs the usual triple-DES and AES encryption standards. What's special is the transmission of the key, a string of random bits used to decipher messages. Computers normally use a random number for the key, producing encryption schemes that could be broken if enough computing power were made available.
"There's a big vulnerability people see, because optical fiber is very easy to tap," says Bob Gelfond, MagiQ CEO, citing one carrier that was finding taps in its Manhattan office "several times a week."
Using a quantum crypto scheme can defend against such taps. In addition to the obvious government and military customers, quantum cryptography is finding interest in the financial sector, for protecting backups or real-time traffic. Another target market would be any industry needing to protect intellectual property -- not just high-tech firms, but businesses such as automotive firms or tire manufacturers, Gelfond says.
But the real market may be the carriers themselves, he notes, simply because they're looking for revenue sources. Quantum cryptography could become a premium service for them. With that in mind, MagiQ is aiming for a price -- around $50,000 to $100,000, depending on features -- that's comparable to other add-ons such as VPN boxes.
Several other companies are working on quantum cryptography, but few appear to be interested in selling a complete system. Swiss firm ID Quantique is trying to commercialize quantum cryptography but so far offers only components such as a photon detector. ID Quantique recently partnered with other Swiss firms to expand its work into a quantum cryptography infrastructure (see Unknown Document 41735).
So, while MagiQ isn't alone in pursuing quantum cryptography, the company's taken a different approach. "Where we started to break new ground was in putting the engineers into the mix, guys who had substantial experience -- Sycamore guys, Tektronix guys," Gelfond says.
MagiQ employs 22, with offices based in New York. Founded in 1999, the company has been powered by roughly $6.9 million in angel funding (see Quantum Crypto Company Launches).
In addition to Navajo, MagiQ is offering a box that only generates the quantum keys, intended as a tool for research outfits and universities.
This is a duplicate post that I copying over from a new story and thread because I think I'm so clever ;-)
-v
---
This saying is getting old on Light Reading, but it is important not to focus too much on the technology, and to look at this from a business perspective.
MagiQ is not selling technology _per_se_.
What these guys are selling is "peace of mind for the ultra-paranoid."
Both the business community and military intelligence/homeland security interests care a lot about information security, risk management, and predictability.
The "ultra-paranoid" have to fear that someone will quietly develop a working quantum computer and put it to work at key cracking, so that computational complexity alone would become ineffective as the basis of encryption.
So the question becomes, isn't $50K or $100K per 100 km a small price to pay in sensitive applications, if that provides insurance against someone quietly developing a quantum computer and cracking your encryption _without_your_knowledge_?
MagiQ doesn't need a mass market to succeed. They can take industry standard components at relatively low cost and build them into a slick quantum key distribution box at a huge mark-up. If they can make it to 1000 units per year, they would have revenues of $50M on a single product, which is fantastic for a start-up.
I predict that they will tap into a rich vein of post- 9/11 paranoia and will make out like bandits!
I'd expect the first users of something like this to be a bank in Manhattan sending their data to the New Jersey backup center in case of fire or similar loss. Those guys probably have already paid for a dedicated optical fiber link, and probably have some big VPN equipment in place too.
I can understand ow a qantum crypto system can be sued to ensure the privacy of a backup sytem like this. However I can see that the bank would also like to prove that the data being backed up came from itself and itself only. To be sure that the data has not been falsified, I do not see how the bank could avoid using a digital signature which relies on some sort of public key system.
I am only trying to understand the basis on which these systems are advocated and am by no means knowledgeable adout these issues. However if the bank must rely on a public key infrastructre to validate its data, why would it not rely on a public key system to keep it private?
Have you ever worked for the NSA? Do you know of people who work for the NSA? Or how about the FBI? CIA? NRO? I suppose if you did you would realize one of two things: a.) The job has no boundries and can never be fully done but has importance to our security. b.) The job can never be completed thus should be terminated.
The reason I think this, is because if you've never worked for any of these agencies how or why would you look at a book written from one persons point of view? Have you wrestled with the concept that he may be biased? Be playing into the people who want to spend their money buying things that prove these Agencies are run by Humans? My point here is don't trust everything you read.
I take your point very serious in one regard. The amount of data to digest is huge! I can't even grasp it it is quite literally amazing. Want to know what else? ITS AMAZING what they do with the data they collect now! But then again everyone wants to point at them and tell them just how horrible they are. That is why the job is thankless. People only know of their flaws and no one gets to know how they have succeeded. But since they exist all these years trust me when I say there are some serious wins in their past. Or so says a little bird.
I really hope I wasn't insulting to you in my prior post? I can't exactly see where I was? In any case I really think the mail system hosted by governments would be a great step. But I'm sure the screaming liberals with privacy issues would jump hand on fist for this... :) Everyone wants the agencies to do more... yet with no more power. Hmm... something has to give.
For your knowledge Quantum computers do exist today. Perhaps not in the classical way, but it's here. And yes the Genie is out of the Bottle. That's why Quantum Security MUST be researched and implemented by our Government. Oh its going to happen trust me. For your information there are a number of projects for the government that will resolve the optics issue. Instead of the mainstay KG-84, 194s, KIV-17s/19s, KG-75s etc. there will be much better gear based solely on optics with quantum key exchange and more (MUCH MORE) But again... the bird could be misquoted.
Yes Jeb, there are many surmountable problems that exist. I for one am a huge advocate for IP_6 as a stepping stone. Personally, I believe that all end-users of any network should be verifiable with location information. (Liberals stand back) Yes that means if you want to participate on the network... Yup your going to need a key. Who will maintain these keys? Why the NSA! :)
I wish I could change your media inspired perception of the NSA... but I won't be able to. Not legally anyway. Just know that while CNN, and the like are out bashing the NSA, the NSA is doing its job and doing it well.
In closing... As a Graduate Student... Hmm. Let me say this. Be humbled and now that truly the only laws placed upon this Universe are the Laws man creates to help provide relief in understanding. Your "I am aware of the practical limitations of building such a system" is the most Arrogant and Ignorant thing repeated throughout history. How many have chocked on those words? I hope I needn't give you examples but let me know if you need a few hundred, I'll be happy to provide you. Should I start with Aristotle? Anyway, Good luck to you and your career. If you do end-up in research I hope your perspective changes for your own sake.
1) I agree with you that the dedicated point to point optical link is not all that cheap, easy, or common place. The units can't be as cheap as a VPN or software on your PC either.
I really think it's a question of what security you want to pay for, and what the value of your data you want to protect is.
I'd expect the first users of something like this to be a bank in Manhattan sending their data to the New Jersey backup center in case of fire or similar loss. Those guys probably have already paid for a dedicated optical fiber link, and probably have some big VPN equipment in place too.
The hard part would be to convince a bank to trust their data to a new security device. Banks don't strike me as first movers on new technology. They like doing their homework and a lot of testing.
If anybody said they were going to put one of these in every home, I wouldn't believe it either just because of the economics.
2) I think the denial of service attack IS the easiest common problem these guys will face.
The good news is, at least you'll know the damn thing is blocked. Anybody using it had better hook in some type of alarm to tell them there's a DOS going on so they can take corrective action, even switching to a new fiber if they have to switch to a new fiber because their fiber company can't fix the break in their old line.
If they don't bother monitoring for even broken lines, and they do keep using their stored keys, I'd almost argue they must not be interested security.
The extra fiber tap put in while the hole was dug shouldn't make any difference. A buddy of mine pointed the technology out to me about a year ago, and I did a quick double take about how things really change depending on your perspective.
If you have 100 photons and put them through a 50/50 splitter, what would you expect?
- Pretty easy, about 50 photons on one tap, and 50 photons on the other.
What happens if you put 100 photons in, one at a time, spaced apart by a few nanoseconds.
- Each one appears at random at one of the two ports. At the end of the process, you should have about 50 on each port.
Boy, that really changed the definition of a 3dB tap, or even 3dB attenuation in a fiber, even though it was the same thing.
So if there is a tap on the line, and the single photons used to send the key are faced with a tap, they get two choices- either take the path to the intended receiver, or take the path to the evesdropper. If the photon doesn't go the intended receiver, it just doesn't get used to make the key.
If the tap is as large as 3dB, you will be able to find it because too many photons are being lost, and if the receiver sees that many photons being lost, it would know something is up, and just like the complete break in the line, that would be like a denial of service attack.
The part I am most skeptical about is the "security is only as good as it's weakest link" part. If it isn't deployed correctly, like someone gets access to the installation area and puts a tap on the "decrypted" side of the box, it doesn't matter what happens on the encrypted side. Like the earlier point, a bank should have a good security policy in place- this box won't help them on that front.
1) The business reason this whole idea (quantum crypto via optics) is a goof is it requires a dedicated point to point optical fiber connection between the "girls". That is hardly the definition of inexpensive, easy or common place.
2) So a backhoe hired by the FBI (or your competitors) comes along and cuts the fiber (oops, crap happens!), then they install their tap somewhere else (while you are stuck using your old key 10,0000 times) and you send your guys out to look at the hole and you fix it and...etc. Fixed and no longer secure...but you can't tell. Looks the same.
Now those realistic problems anyone can understand...
> Don't give me some goddam song and dance > about how I don't understand the product or > the technology you snob. This is exactly the > problem with people like you. You stand up > worshipping some technological boondoggle and > dismiss critics as lesser people.
You are rude. Please try to remain on topic and professional.
It is not your criticism that makes me think of you as I do, it is the manner in which you continue to choose to express it. This includes your quick dismissal of the product as smoke and mirrors intended to swindle suckers.
In post number 24 you accuse Light-Bulb of being personally insulting. I can't find a single instance of him or her being even 1% as insulting as you have been. If you disagree, please point out the insulting part, but keep it professional.
I claimed you did not understand the technology because of your post number 12 in which you state: > Quantum transmission of information is very > lossy and error prone. It might happen sometime. That makes me think you do not understand the principle of quantum error correction. There was no insult intended. That is why I suggested that you do some more search on these related topics, including quantum error correction:
Security of BB84 and B92 protocols: http://www.aip.org/pt/vol-53/iss-11/p22.html
Paper showing Security EPR based quantum cryptography against incoherent symmetric attacks. http://www.iop.org/EJ/abstract/0305-4470/34/35/317
A good primer for other readers on Quantum Key Distribution and Quantum Cryptography. http://www.cs.bris.ac.uk/Research/CryptographySecurity/seminars/Slides/index.html
This is also where you will find a section on quantum error correction codes.
And a little bit about privacy amplification. http://www.ai.sri.com/~goldwate/quantum.html
I hope this information is useful to people following this thread. I don't claim to understand all this material. The point is, that it is not all some crap that you can wave off with a slight of your hand.
If you think this protocol is so easy to break, write a intelligent paper about it and present it at a respectable scientific conference.
You other comment from post number 23: > Your point about operating over temperature > etc... shows that you obviously do not have > any grasp of producing a real product. It's > laughable that you point to BB84 to prove your > point about operating over temperature etc... > shows that you obviously do not have any grasp > of producing a real product. It's laughable > that you point to BB84 to prove your point. > You must have never built a real product in > your life.
Actually, I glossed over it because while telling you to educate yourself, I considered it rude for me to nitpick every mistake in your diatribe.
Since you don't think I can design a product, let's take the opinions of Cisco and Juniper. I hope you consider them to be real product producers, not just smoke and mirror salesmen.
Search for Temperature. Again, a 0C to 40C spec for both units.
And the spec for the QKD contenders mentioned in the main Light Reading article: http://www.idquantique.com/files/spec-qkd.pdf Looks to me like they quote 10C to 30C
http://www.magiqtech.com/press/navajo.pdf Under environmental you will find a 0C to 40C spec.
Do you still consider this to be impossible? Tunable lasers work in NEBS environments where ambient temperatures are tested to 55C or higher to test margins.
I'm guessing that your 75C limit came from a component spec that you worked on in the past.
Personally I think even a 10C to 30C environment is attainable with the advent of modern heating and air conditioning systems that are present in many of today's data centers, but I guess a little margin never hurt.
By the way, I have worked on many real released products; some of which have been tested and passed a 1000 hour life test with -40C to 85C temperature extremes and 100% humidity tests for automotive use with severe shock and vibration testing while operating. Why would you make this attack and then later admit to being a grad student in post number 24? I am assuming that grad students don't have a very deep production design background.
I'll claim that a product should be designed to work in the environment in which it is used, and that starts with the product specification. If the environment is harsher, it should be a different or modified product and specification.
I don't know where you got the idea that I was pointing to BB84 to prove my point about overtemperature not being a problem. If I gave you that idea, I truly apologize. The point was addressed to what I perceived as you not understanding the use of privacy amplification and error correction with the BB84 protocol to increase the security level, and allow the error in the transmission to become immaterial to the security of the key.
Also not stated in the last reply: the obviously low transmission rate is exactly why the usable power of the device is significantly amplified by sending keys with which to send the larger and faster message which is encrypted by that key.
The show in DC has nothing to do with tax dollars (your post #23) any more than having the show in Vegas would make the chance of the product working 18:19 because the odds are in the house's favor. If you don't want to see the MagiQ equipment, go to Switzerland and look at the ID Quantique gear if you want, or maybe you can get an internship at IBM and see a real working system, unless you think IBM is all smoke and mirrors too. (ok I admit you didn't claim their gear was crap too, but I don't think this assumption is going too far out on a limb).
Yes I do think that the fact that you can see one of these systems live and in person at a trade show if you so choose does give some level of credibility to the product category being real, and not just a theoretical lab curiosity. It puts it above all other products that can't get to a working prototype level. The fact that not only one company is claiming a product makes me doubt it even less.
I don't personally think this level of security is infallible, but I think it is a huge level up from the key exchange security present in today's commercial systems.
If you don't agree, great. I really don't care, especially if you continue to be insulting. The post is really for other people who want to read more than just insults.
Point taken and well made until you went off topic and got personally insulting.
I agree that classical computing a lot of this is unecessary. I would like to agree and that it's possible quantum computing may happen (although I don't think in my lifetime but who knows). I would also agree that a quantum computer could break a 128KB key no problem.
I never said though that the government should host the mail servers etc...
My opinion of the NSA is that even if they can break your code, they won't know what to do with the info they find. They are too inept and mired in cloak and dagger buearocracy to be able to be effective against little old me. They can't even handle old time, old fashioned stuff. Read a book by James Bamford (sp?) about how they operate. It's totally laughable what a joke that organization is.
Of course by the time quantum computers are available, the network will be a giant mesh and a mess for people who are trying to eavesdrop on. The problem of finding someone on the network will be as big a job as breaking their codes.
I'm not trying to discourage people from working on it but having done research in this area as a graduate student, I am aware of the practical limitations of building such a system.
Don't give me some goddam song and dance about how I don't understand the product or the technology you snob. This is exactly the problem with people like you. You stand up worshipping some technological boondoggle and dismiss critics as lesser people.
I looked at the product info and basically I my opinion is: this is a lab curiousity and not a real product.
Your point about operating over temperature etc... shows that you obviously do not have any grasp of producing a real product. It's laughable that you point to BB84 to prove your point. You must have never built a real product in your life.
Sure I can invent a theortical way to handle this problem. I can even cheat BB84 with an EPR attack. Sure pal. Better yet, then I suppose you can get around my cheat with BC91. Shortly after that I'll get in my starship and travel at near light speed and travel through a wormhole.
A show in DC? Again another glorious waste of my tax dollars. It's laughable that you think this gives credibility to this product.
Hmm, fritz... perhaps by chance my light is growing dim. After all its been a very rough few years. Then again on the off chance that it isn't, then just maybe I have a point.
Here it is in simple english. You believe that there are a finite number of permutations in a key of (n) size. yes. You are right. You mention that one time pads, or one-way hashs are all guessable. Why yes.. Yes they are. I shall only use one example to prove a point. A system that has been used time and time again throughout human security collection. New language known only by end-points with a TEK,CEK (Whatever lingo your used to seeing) combination on the data stream. Yes... you may be able to break the cipher and decode the stream... but the message is still safe because you haven't the Faintest idea what it means. Thus the Message is safe. That was my point. Sure some of this can be countered with Counter Intelligence but I sure hope you understand what I'm saying. So yes, with any given Encrypted stream there is a chance an attacker can pick the magic key. But who cares if secondary methods are in place. But in our digital world everything speaks the same language. Thus brute force is possible. And with classical computing... I think its VERY safe to say that a 128KB key is unbreakable. At least within a few lifetimes of the known universe. You are incredibly Naive to state the the NSA was not able to decipher... you are the people who believe that carnivore has gone too far. I mean... Maybe the Government should host all Mail servers eh? All corporation would be forced to route through the Governmet servers thus allowing full search and manipulation by the government of mail. With The Quantum computer model keys are not only breakable, they are real-time breakable. Its only years until a few Qubit Systems are available then how far for them to double? 10 years we could have 1024 Qubit systems. Can you understand that? All Codes on earth would be broken in ~1 second. Pretty sexy ehh? We need this technology because someone will have it.
In the context of the discussion about the relative merits of quantum and public key crypto, the use of crypto outside of the protection of messages needs to be considered.
It appears that the main benefit of quantum crypto is the protection of streaming data from eavesdroppers. However there are applications of great importance in which eavesdroppers are part of the normal operation of the system.
Digital signatures are going to be a major aspect of IP tlephony, web apps and other forms of collborative network applciaions. Digital signatuires can be used to create tokens that authorize third parties to act on a user's behalf or to authorize the user to perform some action. For example, they can be given an incoming caller in IP telephony applications to grant them acccess to conferences or other enterpise resources. I believe that SIP uses tokens in this way.
Digital signatures are designed to be presented to third parties. Eavesdropping is a normal part of the process and could be said to the entire point of digital signatures. Quantum does not seem to have any play any part in this.
Public key sytems can easily create and process these signatures. Operators will have to have a scalable facility to process large numbers of signatures. With this requirement, I wonder what real benefit a quantum crypto system could being. It cannot replace the public key infrastructure although the public key infrastructure could perform the functions that quantum crypto does.
In my inexperienced assessment, I find it difficult to see that quantum crypto will find application outside of niche markets.
The blogs and comments are the opinions only of the writers and do not reflect the views of Light Reading. They are no substitute for your own research and should not be relied upon for trading or any other purpose.
To save this item to your list of favorite Light Reading content so you can find it later in your Profile page, click the "Save It" button next to the item.
If you found this interesting or useful, please use the links to the services below to share it with other readers. You will need a free account with each service to share an item via that service.
Digg
Del.icio.us
Reddit
