" Welcome to the Quantum Internet" - By Davide Castelvecch

[reynt0]

On Fri, 22 Aug 2008, Robert J. Hansen wrote:
  . . .
> reynt0 wrote:
>> Wouldn't the claim be:  "cannot listen in on such a
>> transaction between Alice and Bob without affecting the
>> transaction in a detectable way"?
>
> Depends on how pedantic you want to be, and how you define
> 'transaction'.  Frankly, if I were to have proof of an eavesdropper, I
> would consider the transaction to be compromised and I'd scrub it.
> Hence, QM makes it possible to have a key transaction between Alice and
> Bob guaranteed to be free of eavesdroppers.

IIRC, the last time I read in detail about this, there was
bit-by-bit analysis of a full communication A->B, and 
unconfirmed bits were removed from the total exchange,
leaving confirmed bits to constitute the exchanged key
(ie the key offer sent from Alice to Bob).  The cause of
a bit being unconfirmed could be evesdropper "touching"{*}
the bit, circuit noise, etc, so of course DOS is obviously
available by evesdropping on all the bits.  (I'll not try
to state here the reasoning behind the confirmation process,
I'd have to look it up to satisfy myself my memory was good
enough to say anything publicly about it :-) .)

{*}:  I've seen lately, but not reviewed in detail, mentions
of successfully copying photon qbits, by tricky secondary
entanglements, this possibly enabling a functional equivalent
to store-and-forward of photons in their original entanglement 
state.  I think the information was maybe in _Nature_, within
the past couple of  months??, maybe in one of the June 19
articles re quantum stuff (which included an article "The
Quantum Internet" which was listed at the _SN_ article which
started this thread)?