Government Lab Reveals It Has Operated Quantum Internet for Over Two Years

[Sasha]

Government Lab Reveals It Has Operated Quantum Internet for Over Two Years

A quantum internet capable of sending perfectly secure messages has been running at Los Alamos National Labs for the last two and a half years, say researchers

https://www.technologyreview.com/sites/default/files/images/QC%20network.pngOne of the dreams for security experts is the creation of a quantum internet that allows perfectly secure communication based on the powerful laws of quantum mechanics.
The basic idea here is that the act of measuring a quantum object, such as a photon, always changes it. So any attempt to eavesdrop on a quantum message cannot fail to leave telltale signs of snooping that the receiver can detect. That allows anybody to send a “one-time pad (http://en.wikipedia.org/wiki/One-time_pad)” over a quantum network which can then be used for secure communication using conventional classical communication.
That sets things up nicely for perfectly secure messaging known as quantum cryptography and this is actually a fairly straightforward technique for any half decent quantum optics lab. Indeed, a company called ID Quantique (http://www.idquantique.com/) sells an off-the-shelf system that has begun to attract banks and other organisations interested in perfect security.
These systems have an important limitation, however. The current generation of quantum cryptography systems are point-to-point connections over a single length of fibre, So they can send secure messages from A to B but cannot route this information onwards to C, D, E or F. That’s because the act of routing a message means reading the part of it that indicates where it has to be routed. And this inevitably changes it, at least with conventional routers. This makes a quantum internet impossible with today’s technology
Various teams are racing to develop quantum routers that will fix this problem by steering quantum messages without destroying them. We looked at one of the first last year (http://www.technologyreview.com/view/428706/first-demonstration-of-a-quantum-router/). But the truth is that these devices are still some way from commercial reality.
Today, Richard Hughes and pals at Los Alamos National Labs in New Mexico reveal an alternative quantum internet, which they say they’ve been running for two and half years. Their approach is to create a quantum network based around a hub and spoke-type network. All messages get routed from any point in the network to another via this central hub.
This is not the first time this kind of approach has been tried. The idea is that messages to the hub rely on the usual level of quantum security. However, once at the hub, they are converted to conventional classical bits and then reconverted into quantum bits to be sent on the second leg of their journey.
So as long as the hub is secure, then the network should also be secure.
The problem with this approach is scalability. As the number of links to the hub increases, it becomes increasingly difficult to handle all the possible connections that can be made between one point in the network and another.
Hughes and co say they’ve solved this with their unique approach which equips each node in the network with quantum transmitters–i.e., lasers–but not with photon detectors which are expensive and bulky. Only the hub is capable of receiving a quantum message (although all nodes can send and receiving conventional messages in the normal way).
That may sound limiting but it still allows each node to send a one-time pad to the hub which it then uses to communicate securely over a classical link. The hub can then route this message to another node using another one time pad that it has set up with this second node. So the entire network is secure, provided that the central hub is also secure.
The big advantage of this system is that it makes the technology required at each node extremely simple–essentially little more than a laser. In fact, Los Alamos has already designed and built plug-and-play type modules that are about the size of a box of matches. “Our next-generation [module] will be an order of magnitude smaller in each linear dimension,” they say.
Their ultimate goal is to have one of these modules built in to almost any device connected to a fibre optic network, such as set top TV boxes, home computers and so on, to allow perfectly secure messaging.
Having run this system now for over two years, Los Alamos are now highly confident in its efficacy.
Of course, the network can never be more secure than the hub at the middle of it and this is an important limitation of this approach. By contrast, a pure quantum internet should allow perfectly secure communication from any point in the network to any other.
Another is that this approach will become obsolete as soon as quantum routers become commercially viable. So the question for any investors is whether they can get their money back in the time before then. The odds are that they won’t have to wait long to find out.


Ref:arxiv.org/abs/1305.0305 (http://arxiv.org/abs/1305.0305?utm_source=feedly):Network-Centric Quantum Communications with Application to Critical Infrastructure Protection


quote:http://www.technologyreview.com/view/514581/government-lab-reveals-quantum-internet-operated-continuously-for-over-two-years/

[Brutus]

this seems awesome

[ckaihatsu]

Actually it seems redundant, given the security of the one-time pad itself:





In cryptography, the one-time pad (OTP) is a type of encryption which has been proven to be impossible to crack if used correctly.




If the key is truly random, as large as or greater than the plaintext, never reused in whole or part, and kept secret, the ciphertext will be impossible to decrypt or break without knowing the key.[1][2]




http://en.wikipedia.org/wiki/One-time_pad

[Sea]

Actually it seems redundant, given the security of the one-time pad itself:Let's say our good friends Alice and Bob have some important poop to pass down the pipeline. Alice sends her one-time pad down the pipe and waits for Bob to acknowledge delivery. Mysterious Mr. X cracks a hole into the pipe and sticks his grubby little hands in, copying the pad and sending the original on its way to Bob unchanged. Bob gets it and lets Alice know. When it comes time for Alice to send her top secret spy stuff to Bob, Mysterious Mr. X has a peek at it and Bob and Alice are none the wiser.

Obviously it's not this simple in real life, but your message is only as secret as your OTP. When you're using encryption, you're not adding any security unless you can safeguard your pad / password / whatever better than you could safeguard the plain unencrypted data.

Consider an encrypted hard drive. You can't really trust hundreds of gigabytes of unencrypted data to remain secret should your laptop be stolen. I can't, anyway. By using disk-level encryption, you're instead trusting the integrity of a smaller, more manageable passphrase. If you write it down, you're placing your trust in a snippet of paper. If you memorize it, you're placing your trust in your own soundness of mind. In either case you're also trusting the ability of whatever encryption algorithm to successfully use a small secret to secure a large bank of data, which can be quite a problem in itself.

This is where the one-time pad comes in. Having a one-time pad as large as the data you want to encrypt eliminates that last one and only that last one. From a practical standpoint this is largely self-defeating. Instead of guarding 12345 bytes of secret data, you now have to guard 12345 bytes of pad. Now that your data is encrypted in a mathematically flawless manner, you can let it go over insecure channels out in the wild. But guess what? You still have 12345 bytes of sensitive information to guard. Sure, now what you have left to keep secret isn't human-readable, but that's not security, it's obscurity.

The ability to guarantee the integrity of your pad is a huge step in making full-length one-time pads practical. As I have already outlined, practical is the key word here. Now Bob can tell if the pad that Alice sent has been tampered with, and if it has been, refuse the use of whatever communications channel that Alice and Bob were using.

If the pads are random like they should be, how will Bob be able to tell if it has been tampered with? Perhaps a header like 123456 or ABCDEFG before the pad, so that if the header doesn't match up the message could be refused. Of course, Mysterious Mr. X could ignore reading the header if he knows how long it is and its location, therefore leaving it unchanged, but the relatively short header could be pre-agreed upon by Alice and Bob, both of them having memorized its location in the message and its contents.

That way, you get the unbreakable security of a full-length one-time pad and the practicality of a simple passphrase system. Would this work? No clue. Honestly I'm just rambling at this point.

[ckaihatsu]

Let's say our good friends Alice and Bob have some important poop to pass down the pipeline. Alice sends her one-time pad down the pipe and waits for Bob to acknowledge delivery.


This would be an improper protocol to use at all, for the reasons you proceed to outline.

Best would be in-person contact for any future person-to-person secure communications, though the public key method has a built-in authenticating step for any given arbitrary parties:


http://en.wikipedia.org/wiki/Diffie%E2%80%93Hellman_key_exchange





Mysterious Mr. X cracks a hole into the pipe and sticks his grubby little hands in, copying the pad and sending the original on its way to Bob unchanged. Bob gets it and lets Alice know. When it comes time for Alice to send her top secret spy stuff to Bob, Mysterious Mr. X has a peek at it and Bob and Alice are none the wiser.

Obviously it's not this simple in real life, but your message is only as secret as your OTP. When you're using encryption, you're not adding any security unless you can safeguard your pad / password / whatever better than you could safeguard the plain unencrypted data.

Consider an encrypted hard drive. You can't really trust hundreds of gigabytes of unencrypted data to remain secret should your laptop be stolen. I can't, anyway. By using disk-level encryption, you're instead trusting the integrity of a smaller, more manageable passphrase. If you write it down, you're placing your trust in a snippet of paper. If you memorize it, you're placing your trust in your own soundness of mind. In either case you're also trusting the ability of whatever encryption algorithm to successfully use a small secret to secure a large bank of data, which can be quite a problem in itself.

This is where the one-time pad comes in. Having a one-time pad as large as the data you want to encrypt eliminates that last one and only that last one. From a practical standpoint this is largely self-defeating. Instead of guarding 12345 bytes of secret data, you now have to guard 12345 bytes of pad. Now that your data is encrypted in a mathematically flawless manner, you can let it go over insecure channels out in the wild. But guess what? You still have 12345 bytes of sensitive information to guard. Sure, now what you have left to keep secret isn't human-readable, but that's not security, it's obscurity.

The ability to guarantee the integrity of your pad is a huge step in making full-length one-time pads practical. As I have already outlined, practical is the key word here. Now Bob can tell if the pad that Alice sent has been tampered with, and if it has been, refuse the use of whatever communications channel that Alice and Bob were using.


I appreciate the rundown, but secure communications are a different matter from that of secure backups of permanent data. The idea behind the OTP is that it's limited to a *message* size, and that it's disposable, per message.





If the pads are random like they should be, how will Bob be able to tell if it has been tampered with? Perhaps a header like 123456 or ABCDEFG before the pad, so that if the header doesn't match up the message could be refused.


The mutually accepted protocol could use hashing, for a "heads-up", 'pre-verified' step, if you like.





Of course, Mysterious Mr. X could ignore reading the header if he knows how long it is and its location, therefore leaving it unchanged, but the relatively short header could be pre-agreed upon by Alice and Bob, both of them having memorized its location in the message and its contents.

That way, you get the unbreakable security of a full-length one-time pad and the practicality of a simple passphrase system. Would this work? No clue. Honestly I'm just rambling at this point.


If you're interested in this kind of stuff, here's yet another approach to the problem:


http://en.wikipedia.org/wiki/Chaffing_and_winnowing