Government Lab Reveals It Has Operated Quantum Internet For Over Two Years | Hacker News Reader

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

technologyreview.com

112 pointsmtgx35 comments

Calling something that can't do routing an "Internet" is pure hyperbole.

Without quantum routers, you can only do quantum key exchange with the router you're directly connected to. That router can only do quantum key exchange with things it is directly connected with. If any router on the communication path is untrustworthy... too bad.

If you want quantum crypto to be practical, you need routers that can redirect messages without measuring or copying them. That way you don't need to trust the network; only the endpoint you're talking to. (One possibility would be to setup dedicated circuits ahead of time, like the original telephone system, before qbits started being exchanged.)

Not sure if your beef is with the capital I or with the term "internet". Switching packets between multiple point-to-point networks is certainly within the realm of internetworking although I would agree that routing at layer 3 would come much closer to agreeing with a stricter definition of an internet.

In any case, it isn't really hyperbole unless you think they are drawing a comparison to the Internet (capital I).

Agreed. "Quantum LAN" would be a more appropriate, if less exciting, term.

> measuring or copying

That was redundant :)

I feel like the best feature of our our internet is the lack of a central hub like the network described in this article, and that "internet" isn't really an appropriate term here.

edit: super awkward wording.

internet = generic term for a set of interconnected networks Internet = proper noun describing one particuarly well-known internet

I may have misread the article but my understanding was that they have built just a single network, is this correct?

Sounds like a single layer 3 network but multiple layer 2 data links. Usually 'internet' is reserved for networks connected by a layer 3 device but a layer 2 device still has to make 'routing' decisions between the data link layers, which is usually called 'switching' and not routing.

I'm not sure if there is a common term for an aggregate collection of switched data link segments, perhaps "layer 2 network"?

In any case, using "internet" to refer to connected layer 2 data links isn't a particular strong abuse of the term in my mind.

They state that it's impossible to secretly listen in (because that would involve tampering with the data), but then also state that the hub works by converting to classical bits and then retransmitting.

What prevents an attacker from inserting a device that converts to classical bits before retransmitting?

Exactly. The hub itself must be secure and trustworthy for an entirely 'secure' communication route, this won't work across any plain-jane network due to that weakness.

The hub is where the NSA grabs their copy of the key.

I believe the retransmitted quantum packet would be necessarily different from the original. So the receiving node notices that the packet has been modified, but because it expects the (secure) hub to modify the original, all is well.

But how can it tell the difference between the hub's modifications and an attacker's modifications?

it can't, that's why you need a trusted hub

clients can only tell that the message they received from the hub was actually from the hub and was not observed by anyone but the hub. there's no way to tell a message from client A to client B wasn't tampered with or observed before the hub delivered it

So if you're supposed to trust all hops on your message, quantum cryptography has no real appreciable improvement over standard digital communications (where the same problem exists).

Well, you don't have to worry about anyone hijacking a cable. Rather big improvement there. Although not without cost...

These researchers come from secretive Los Alamos National Labs, too. If LANL is willing to divulge this work, what do ULTRA-secretive places like Fort Meade have going?

This sounds like a job for Alex Jones! But, seriously though, we know about this because they published a research paper. It's not like this is leaked, classified information.

That's what he's saying. If this is the NSA telling us what the s-boxes for DSA are, what's the equivalent to the DSA understanding differential cryptanalysis when designing the s-boxes?

i think you mean DES not DSA.

Whoops, my bad.

If they told you, they'd have to... oh never mind.

Here's the link to the researchers' paper on arXiv.org http://arxiv.org/abs/1305.0305

So it's a bunch of point-to-point links between computers in government funded research labs, approximately where the capital-I Internet was in 1970?

Am I understanding this correctly, that a quantum internet prototype has been in operation with at least the same level of performance as modern networks? This is fantastic!

There is no routing involved, it is only transmitting message from A to B on a single fiber. I think the term "Internet" in this context is a bit of an overstatement.

I've no experience with networking. How complex is it to incorporate routing?

They talk about the problem with it so far in the article:

"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"

So to incorporate routing, we're going to have to get a bit more clever.

Well, the way that immediately comes to mind would be to send two packets per transmission. One that is point-to-point, which contains the routing information, and one that is meant to be routed without being read, but there must be something more to it as this solution seems too simple to have not been thought of.

Well they do mention that they "solved" this problem:

"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."

With a little more reading, it seems like the issue they face now is the reliance on a secure central hub.

TFA talks about it a bit, but your biggest problem is multiplexing/demultiplexing, since that requires you to essentially create a MITM attack (i.e. read the data at a 3rd point besides the sender/reciever), which kind of defeats the purpose of quantum computing. They mention a hub based system, but again, TFA talks about scaling which can be very hard to do when you're essentially relying on a single point of failure that has to scale with the entire network.

Your question is basically what the entire article is about. Short answer: difficult when dealing with a system that by design is not supposed to read any part of the data.

Impossible with today's technology = extremely difficult.

Routers have to read the data in order to know where it is supposed to be routed to. Quantum routing isn't allowed to do that. That's a difficult problem.

edit: apologies for duplicating previous answers, should've refreshed the comments page before posting.

Hypothetically impossible. Routing requires understanding where the packet is to go, combining it with other packets going through the "router" and then sending those along to their destination in an efficient way. To do this inspection of the packets, you have to get information from it in some way. Our understanding of quantum physics doesn't allow this, so trying to inspect the packet in any way will create a false positive that the encryption has been broken. There's no way(again, that we know of) that can tell a false positive from a real positive in this case.

Or you can just transmit routing info in a side channel that runs on conventional technology.

The hub is certainly doing switching if not routing between multiple point-to-point networks. If this is layer 3 routing than "internet" is a fair term. Even if it is layer 2 switching, "internet" isn't too far off the mark.

On the other hand "Internet" is not correct at all since that is one particular "internet" distinctly different than the internet described in the article.

Reply on news.ycombinator.com