Advance opens door for secure quantum applications without specialized infrastructure
So they’ve shown they can send light over a cable designed to transfer light.
The impressive thing is of course managing to get one specific photon to one specific location. Still, what benefits does that have over the standard encoding?
I guess this technique might have a lower error rate and higher distance, because it’s binary by nature with no quantization needed. But you don’t need the quantum entanglement part at all for this.
Edit: Reading is hard! This is indeed exciting for security. I wonder how it fairs against a very powerful MitM though.
It’s physically impossible to intercept an entangled photon without disrupting the entanglement. The act of observing the photon collapses the quantum uncertainty of it’s state, so even the most sophisticated MitM attempt is going to immediately break the link.
But can you detect the link being broken by someone other than your intended communication partner?
If A sends a particle to B, couldn’t M intercept A’s particle and send a different particle to B? Kind of like intercepting Diffie Hellman. A and B will both share some information with M, but not with each other.
M cannot replicate the particle after they read it, so A and B will detect M’s attempt when they compare results.
The same as classical one-way encryption, it only works through authenticated channel tho. It’s not magic, you have to have some kind of pre-existing secret or rely on third-party authentication
But can you detect the link being broken by someone other than your intended communication partner?
Yes, because breaking the entanglement destroys the link between the photons received at either end.
Observing an entangled photon requires extremely precise timing, the lightspeed lag on the line has to be known down to the nanosecond to ensure that the photon received is paired with the photon at the other end. Even if a MitM wanted to try retransmitting the quantum states it observes on the line, they wouldn’t be able to do so without introducing enough lag to desync the connection.
Alternatively, if M tried sending their own random data in sync with the expected timing, then the bits received by B would only have a 50-50 chance of matching the bits sent from A. Any encryption based on that data would almost immediately begin to suffer a 100% error rate.
