- Name: Christian Kurtsiefer (NUS [Singapore])
- Email address: firstname.lastname@example.org
- Employer (if any), University/College: Centre for Quantum Technologies/National University of Singapore
- Shipping address and instructions - telephone number
- any special instructions
- power adapter: European Standard (240V), but we have enough 12Vdc supplies
- desired keyboard layout : US English
- Quantity of machines desired: 3
- Report on their experiments at the ccc
What is this?
This is a fun project to play with quantum information in a reasonably serious way.
Getting the 'refute local realism' QKD toolchain to work using only 2 XOs and 60kg of equipment.
Attaching a pair of telescopes to a source of entangled photon pairs, capturing a stream of measurement data from said pairs on two machines and use the results to generate a secret key or violate a Bell inequality.
The machines need to run a kernel driver for a data acquisition system, FFT and other correlation algorithms, NTPD or something equivalent, and software to remotely correct errors etc. We adapt our user interface to watch the key generation and the violation of a Bell inequality from the measurements.
This project started spontaneously on the 24C3 as a jam session, with no deep thoughts behind other than enjoying to do something crazy.
We are a physics research group  working on quantum key distribution and other aspects connected with entanglement between different physical systems, one of the funny things which comes out of quantum mechanics. We have a reasonably compact physical hardware providing entangled photon pairs, which we brought along to the 24C3 with a tow telescopes and photodetectors, where out of a series of measurement results finnaly a secret key can be generated between two measuring parties.
Apart for the application of generating a secret key, one can demonstrate nicely that a whole class of theoretical alternatives to quantum mechanics are not valid by violating the so-called Bell inequality for a combination of measurement outcomes. As far as experiments probing fundamental aspects of quantum physics go, this is really simple.
Usually we run the reasonably involved correlation software for such measurements on two normal laptops or desktops, which connect to some self-knitted hardware for timing measurements via USB to the PC for processing raw data rates of 5 to 30 Mbyte/sec.
What happened on the 24C3
Somehow we liked by the very neat and lean hardware of the XO machines, and thought it would be cute to play with a fundamental physical effect on a machine which appears like a toy but is nifty enough to do serious processing. With support form the oplc people there and help from friendly kernel programmers at the meeting we ported the large bandwidth device driver for our hardware and all the other components of the correlation software to the olpc. That worked out remarkably smooth, and the really critical components started to run on it nicely.
We got up to the point where we could identify photon pairs for a key exchange, but parts of our software generated probably tons of logging info flooding the flash memory, and we stopped when the wifi at this place became a bit flaky, perhaps due to a half watt monster base station next to our camp.
It would be really nice to finish this project up cleanly. Not sure if this blurb here fits the format, but that's roughly what we have in mind to do that:
Requirements We would need two xo machines similar to those you had at the 24C3. The rest of the hardware for photodetection etc. is here in Singapore.
Demo We run a key distribution demo, and demonstrate the violation of a Bell inequaltiy over some distance (how far does your wifi go?) to demonstrate that the XO can be used for playing with quantum information.
Time scale The whole thing will probably take about 1-2 weeks.
Open source All the code we use is GPLed and hopefully reasonably standard linux. But we do not really expect that every child will have an entangled photon pair source anytime soon....
We think that your hard- and open software is really well-suited for many lab and field data acquisition jobs. If we get a quantum crypto experiment working on them after a beer or two in a reasonably short time, people get lots of other measurement tasks done with it as well. And it hardly uses any power and is not overly picky in what it gets (Who likes to carry around batteries or generators....)
You probably find interest in all smaller scale research and teaching labs, also as OEM units just with the screen and motherboard, without the keyboard/touchpad. We could already think of a few places in our lab where we would use them if we could get hold of them for a reasonable price.
Would it work for you that you sell the barebone units as OEM devices to perhaps industrial or other customers to run a get x / give y scheme for the full PCs?