Shmoocon 2016: Computing In A Post Quantum World

January 16, 2016 by Brian Benchoff 50 Comments

There’s nothing more dangerous, so the cryptoheads say, than quantum computing. Instead of using the state of a transistor to hold the value of a bit as in traditional computers, quantum computers use qubits, or quantum information like the polarization of a photon. According to people who know nothing about quantum computers, they are the beginning of the end, the breaking of all cryptography, and the Rise of the Machines. Lucky for us, [Jean-Philippe Aumasson] actually knows a thing or two about quantum computers and was able to teach us a few things at his Shmoocon talk this weekend, “Crypto and Quantum and Post Quantum”

This talk is the continuation of [Jean-Philippe]’s DEF CON 23 talk that covered the basics of quantum computing (PDF) In short, quantum computers are not fast – they’re just coprocessors for very, very specialized algorithms. Quantum computers do not say P=NP, and can not be used on NP-hard problems, anyway. The only thing quantum computers have going for them is the ability to completely destroy public key cryptography. Any form of cryptography that uses RSA, Diffie-Hellman, Elliptic curves is completely and totally broken. With quantum computers, we’re doomed. That’s okay, according to the DEF CON talk – true quantum computers may never be built.

The astute reader would question the fact that quantum computers may never be built. After all, D-Wave is selling quantum computers to Google, Lockheed, and NASA. These are not true quantum computers. Even if they’re 100 Million times faster than a PC, they’re only faster for one very specific algorithm. These computers cannot simulate a universal quantum computer. They cannot execute Shor’s algorithm, an algorithm that finds the prime factors of an integer. They are not scalable, they are not fault-tolerant, and they are not universal quantum computers.

As far as true quantum computers go, the largest that has every been manufactured only contain a handful of qubits. To crack RSA and the rest of cryptography, millions of qubits are needed. Some algorithms require quantum RAM, which nobody knows how to build. Why then is quantum computing so scary? RSA, ECC, Diffie-Hellman, PGP, SSH and Bitcoin would die overnight if quantum computers existed. That’s a far scarier proposition to someone hijacking your self-driving car or changing the display on a smart, Internet-connected thermostat from Fahrenheit to Celsius.

What is the verdict on quantum computers? Not too great, if you ask [Jean-Philippe]. In his opinion, it will be 100 years until we have a quantum computer. Until then, crypto is safe, and the NSA isn’t going to break your codez if you use a long-enough key.

Back To Basics: What’s The Deal With Magnets?

November 20, 2015 by Dan Maloney 82 Comments

I consider myself a fairly sharp guy. I’ve made a living off of being a scientist for over 20 years now, and I have at least a passing knowledge of most scientific fields outside my area. But I feel like I should be able to do something other than babble incoherently when asked about magnets. They baffle me – there, I said it. So what do I do about it? Write a Hackaday post, naturally – chances are I’m not the only one with cryptomagnetonescience, even if I just made that term up. Maybe if we walk through the basics together, it’ll do us both some good understanding this fundamental and mysterious force of nature.

Continue reading “Back To Basics: What’s The Deal With Magnets?” →

Quantum Cryptography In-band Attack

October 6, 2008 by Eliot 10 Comments

Quantum cryptography is an emerging field, but low install base hasn’t kept researchers from exploring attacks against it. It’s an attractive technology because an attacker sniffing the key exchange changes the quantum state of the photons involved. All eavesdroppers can be detected because of this fundamental principal of quantum mechanics.

We’ve seen theoretical side-channel attacks on the hardware being used, but had yet to see an in-band attack until now. [Vadim Makarov] from the University of Science and Technology in Trondheim has done exactly that (Internet Archive). Quantum key distribution systems are designed to cope with noise and [Makarov] has taken advantage of this. The attack works by firing a bright flash of light at all the detectors in the system. This raises the amount of light necessary for a reading to register. The attacker then sends the photon they want detected, which has enough energy to be read by the intended detector, but not enough for the others. Since it doesn’t clear the threshold, the detectors don’t throw any exceptions. The attacker could sniff the entire key and replay it undetected.

This is a very interesting attack since it’s legitimate eavesdropping of the key. It will probably be mitigated using better monitoring of power fluctuations at the detectors.

[via I)ruid]

More Defcon 16 Events Announced

August 2, 2008 by Eliot 1 Comment

Defcon keeps announcing more and more interesting events for next week’s conference. A free workshop is planned for the soon to be released DAVIX live CD. DAVIX is a collection of tools for data analysis and visualization. They’ll be running through a few example packet dumps to demonstrate how the tools can help you make sense of it all. [Thomas Wilhelm] will be driving out from Colorado Springs in his Mobile Hacker Space. He’s giving a talk Sunday, but will be giving presentations a few hours every day at the van. Some researchers from NIST will be setting up a four node quantum network and demonstrating some of the possible vulnerabilities in the system. Finally, as part of an EFF fundraiser, Defcon will feature a Firearms Training Simulator. Conference attendees will participate in drills designed to improve their speed, accuracy, and decision making skills.