'The cloud' is so cliché Microsoft even has housewives using it in commercials.  All it basically means is that you can have a new way for your data to be hacked and stolen.  Unless the company storing it goes out of business and you made no backups.  Plus, that's not really a cloud, it's stored in a solid medium, it just happens to be outside your house.

But true cloud storage is on the horizon. 

Light signals have been stored as patterns in a room-temperature atomic vapor by scientists at the Joint Quantum Institute. They stored two letters of the alphabet in a cell filled with rubidium (Rb) atoms. 

It's not just cool cloud storage, it is also the first stored and replayed atomic movie, if storing and replaying two separate frames a few micro-seconds apart qualifies as a feat of cinematography. 

It even has a soundtrack. 

Catchy?  Like the lyrics?  Read the arXiv paper and sing along.  

The quantum device itself is not all that quantum.  It's 7 inches long but gradient echo memory (GEM) needs a lot of room.

The image is stored in this extended way, by being absorbed in atoms at any one particular place in the cell, depending on whether those atoms are exposed to three carefully tailored fields: the electric field of the signal light, the electric field of another "control" laser pulse, and a magnetic field (adjusted to be different along the length of the cell) which makes the Rb atoms (each behaving like a magnet itself) precess about. When the image is absorbed into the atoms in the cell, the control beam is turned off. Because this process requires the simultaneous action of two particular photons---one putting the atom in an excited state, the other sending it back down to a slightly different ground state---it cannot easily be undone by atoms subsequently randomly emitting light and returning to the original ground state.

That's how the image is stored. Image readout occurs in a sort of reverse process. The magnetic field is flipped to a contrary orientation, the control beam turned back on, and the atoms start to precess in the opposite direction. Eventually those atoms reemit light, thus reconstituting the image pulse, which proceeds on its way out of the cell.

Having stored one image (the letter N), the JQI physicists then stored a second image, the letter T, before reading both letters back in quick succession. The two "frames" of this movie, about a microsecond apart, were played back successfully every time, although typically only about 8 percent of the original light was redeemed, a percentage that will improve with practice. According to Paul Lett, one of the great challenges in storing images this way is to keep the atoms embodying the image from diffusing away. The longer the storage time (measured so far to be about 20 microseconds) the more diffusion occurs. The result is a fuzzy image.

Paul Lett plans to link up these new developments in storing images with his previous work on squeezed light. "Squeezing" light is one way to partially circumvent the Heisenberg uncertainty principle governing the ultimate measurement limitations. By allowing a poorer knowledge of a stream of light---say the timing of the light, its phase---one gain a sharper knowledge of a separate variable---in this case the light's amplitude. This increased capability, at le ast for the one variable, allows higher precision in certain quantum measurements.

"The big thing here," said Lett, "is that this allows us to do images and do pulses (instead of individual photons) and it can be matched (hopefully) to our squeezed light source, so that we can soon try to store "quantum images" and make essentially a random access memory for continuous variable quantum information. The thing that really attracted us to this method---aside from its being pretty well-matched to our source of squeezed light---is that the ANU group was able to get 87% recovery efficiency from it - which is, I think, the best anyone has seen in any optical system, so it holds great promise for a quantum memory."

Citation: Quentin Glorieux, Jeremy B. Clark, Alberto M. Marino, Zhifan Zhou, Paul D. Lett, 'Temporally multiplexed storage of images in a Gradient Echo Memory', Optics Express, Vol. 20, Issue 11, pp. 12350-12358 (2012) DOI: 10.1364/OE.20.012350. Preprint on arXiv.