lcs Mixmaster Remailer wrote:
> Quantum cryptography will be of little practical value for the average
> person. That's because you need to get photons unchanged from one
> person to the other. This requires either a line of sight or a fiber
> optic cable, neither of which is likely to be available.
If they became useful, fibre optic cables would be made available. It is
probably the future (I know, I know, we've been saying this for 20 years
& it hasn't happened yet) but if one fibre has a significant fraction of
the the bandwidth of the entire sky it has to be the way to go.
Anyway - who says radio isn't transmitted by photons?
Satellites communicate by line-of-sight, both with each other and with
endpoints. Laser comms in space?
It explains the Fermi paradox anyway - they are out there but they live
on iceballs in the Kuiper and Oort and communicate by store-and-forward
through tight-beam lasers using quantum cryptography techniques to
error-check the messages over those distances... so we never get to
intercept their comms. Travel from star to star by a long series of
short hops from chilly blob to chilly blob. I have seen the future of
interstellar communications and it looks a lot like Usenet That's what
happened to S****r A***c you know - when his stuff got out to Alpha
Centauri the aliens came and got him.
> Quantum computers allow fast search for symmetric ciphers like DES
> or AES. The effect is essentially to halve the key size. A 128 bit key
> attacked by a QC becomes as strong as a 64 bit key would be attacked by
> conventional computers. The new AES standard provides for 256 bit keys.
> These will still provide 128 bits of strength against quantum computers,
> making them practically invulnerable. So QCs will provide no significant
> problems against symmetric ciphers once AES is in widespread use.
>
> Quantum computers also allow fast factoring and finding discrete logs,
> essentially destroying the principles behind the most widely used
> public key systems. This uses Shor's algorithm, which works by finding
> the period of a sequence. The recent IBM announcement was apparently
> an implementation of just this algorithm for a 5 bit QC.
>
> Hence it will be necessary to scale up the QC from 5 bits to 1024 bits
> or more. This will take years of work and no one knows if it will be
> possible. If it happens, people will have to switch to keys larger than
> the largest quantum computers, which will probably be a losing battle;
> or they will have to use the more obscure, less efficient and possibly
> less secure public key alternatives. No doubt if large QCs appear on
> the horizon we will see considerably more cryptographic effort put into
> developing and establishing the security of alternative methods for PKC.
Or we just get a lot of people who are good at sums to work on
non-paralellisable algorithms, where the output of stage n must be known
before n+1 can be set up. The opposite of what they are doing now of
course. Though who knows what the NSA are up to - maybe if they believe
all this QC stuff they have been paying people for years to work out
deliberately inefficient, unoptimisable algorithms. It's a living.
Ken (& not the College)
