In the old days, you had to get security clearance to access a supercomputer. Now, just reach inside your pocket. Cloud computing gives everyone the ability to access the power of an ultra-available, distributed network of servers. However, it’s exactly that ultra-availability that concerns skeptics of this “third platform” technology. People worry that by processing their data through a hosting service’s system, either the host or one of its customers could take a peek.
Hence, hosting companies have started offering private and hybrid cloud services. But people remain hungry for stronger security mechanisms within the public cloud. A new discovery wants computing to take a quantum leap.
As Jacob Aron of New Scientist reports, quantum state manipulation can optimize the security of distributed virtualization.
Two Established Cloud Security Ideas
Two ways to bolster the security of the public cloud have emerged in recent years, says Aron. One is homomorphic encryption.
“A homomorphic encryption scheme is a crypto system that allows computations to be performed on data without decrypting it,” explains Andy Greenberg of Wired. “A homomorphically encrypted search engine, for instance, could take in encrypted search terms and compare them with an encrypted index of the web.”
Put another way, homomorphic encryption is a sophisticated algorithmic approach that allows you to submit information to the cloud server completely scrambled rather than sunny-side-up. Without using the approach, your virtual machine would ordinarily be incapable of manipulating the encrypted information.
The other way is quantum manipulation – computing that uses nuclear magnetic resonance to enact quantum-mechanical processes, such as superposition and entanglement, on molecules from ordinary liquids. In 2012, Stefanie Barz and her team of researchers at the University of Vienna presented methods to alter quantum states in order to render a quantum machine incapable of “seeing” information as it runs tasks.
Though strong in terms of security, each of these mechanisms has practical weaknesses: homomorphic encryption is extraordinarily resource-intensive, and quantum manipulation involves substantial client-server interaction – spending more time talking the talk than it does walking the walk.
Realizing that her revolutionary method was too excessive on the call and response, Barz developed a new version to scale back the chatter.
Okay, Let’s Be Logical
For her new research project – released January 27, Barz’s team created a simulation of cloud processing in miniature – with just one bit. Server and client “computers” were set up in rooms that were approximately 50 yards apart, mimicking the relationship between cloud hosting services and those who use them (including companies with cloud systems and everyday Internet users).
The server created cobits, which are essentially rudimentary quantum photons. The typical quantum unit is a qubit. Here is how those two units compare:
- qubit (not to be confused with Q*bert, who is also in the news) – zero, one, or a blend of the two with broad possibilities, the last option allowed by quantum superposition
- cobit – zero, one, or an exact superposition; a specification that makes this unit less sophisticated than the qubit but also more practically useful (i.e., less client-server back-and-forth).
Barz commanded the server to send cobits to the client, the device in the other room at the University of Vienna that microcosmically represented a person wanting to experience hyper-secure cloud. The client was an optical construction that carried out a simple logical processing scheme via an exclusive-or gate, a.k.a. a XOR gate.
Although the client was incapable of producing cobits, it was able to manipulate the cobit’s state into an encrypted version of the data exercise it wanted to conduct.
“The server [could] then perform a quantum measurement on the cobit in order to run a more powerful computation, called a NAND [a.k.a. “not-and”] gate,” Aron explains. “In theory, using a NAND gate repeatedly could perform any computation, making this a neat demonstration of the scheme’s potential.”
Why It Works & Practical Challenges
In this scenario, because the client machine encrypts the data, the server does not understand its own outputs. They are simply returned to the client, which is the only Bilbo Baggins in the land with the secret Cracker Jack decoder ring to read the information.
In this way, a user is able to benefit from the cloud’s supercomputer potential while enforcing a cloak around the data.
Although using cobits better suits quantum computing to the real world, ID Quantique head Gregoire Ribordy is not sure that the model developed in Vienna would actually work in day-to-day operations. The hole in the puzzle, says Ribordy, is that (Aron paraphrase) “any real-world client device will likely be capable of computing a NAND gate.”
Scientific research into quantum manipulation is helping to build the cloud of tomorrow. Today, standardized protections of public cloud are verified by SSAE-16 auditing. Talk to us about hybrid and private solutions as well.
By Kent Roberts