This paper presents a working implementation of a quantum key 
distribution, QKD, point-to-point link. The architecture has two 
channels: one is for transmission of quantum key material and the other 
one carries all message traffic. The guarantee of a quantum channel is 
that if an eavesdropper monitors the channel, this monitoring will 
cause measurable disturbance. The authors describe physical layer 
implementation of QKD that involves sending a single photon, modulated 
to one of four phases, and detecting this photon phase on the 
receiver's end. To interpret the result correctly, both the sender and 
the receiver must use compatible bases of modulation. Because of lost 
photons, transmission interruptions and incompatible modulation bases, 
losses in qubits' transmission can be significant. Therefore, the 
authors develop a multi-step process that attempts to clean up 
mistakes, correct errors, amplify privacy by eliminating exposes bits 
and finally authenticating both parties in!
  a!
 transmission. QKD can become an extension of IPsec protocol, so the 
authors demonstrate changes that are necessary to implement such an 
extension. Finally, future work includes extending the QKD network to 
provide redundancy and exploring theoretical properties of quantum 
cryptography. This is crucial work because in the word of the authors: 
"it is now clear that quantum cryptography is feasible in practice -but 
the question still remains as to whether it's feasible in theory."

Pros:
- Quantum cryptography can be shown to be provably secure
- Idea of a quantum network is interesting
- The paper presents actual implementation of a quantum link
- Quatum cryptography moves the whole field to a provably secure model 
of protection

Cons:
- Authorization is still a problem
- Work is based on assumptions about eavesdropping
- Excessive focus on IPsec

Strong Accept - 1
Accept        - 8
Reject        - 2
Strong Reject - 0

Michael