UPDATE, March 10, 2018: computing technology update, Google’s Bristlecone Quantum Processor.
Throughout the week of April 11th, 2016 Stanford held is annual affiliates Computer Forum on the campus. Participation in the forum is available to affiliate members. If your interested to be an affiliate send a note to me, see About page. Stanford security forum is a great place to unplug from the day-to-day business and consider broader security challenges. The campus is beautiful and the projects are interesting. Attending the forum is always uplifting, I usually meet leaders from industry I know, university staff, and I always learn something new from their research.
The forum is a week long but attendees can sign up for individual days depending up interests. I attended 2 days of the week long forum. Monday was dedicated to security. Thursday was dedicated to IoT. Research projects and themes change from year to year. This year cryptography and IoT where the broad themes. Full media from the week long forum trails the post.
A Few Thoughts or Impressions
Following are some of the more important points I learned or points that captured my interests, not in any particular order of importance.
Why are quantum computers fast?
Traditional computers process information in bits. A bit is either “on” or “off”, a 1 or a 0 respectively but quantum computers also provide an Amplitude property associated with each quantum bit. Remember Schrödinger’s Cat? The cat was in a Superposition of States where the cat is both alive and dead. Amplitude is the measurement of the superposition which is the probability the cat is in one state or the other. A point of some utility is that amplitude is not a simple percentage but instead is a complex number. The the value combined with the amplitude of the bit form a quantum computational unit known as the Qubit. In a traditional computer, increasing the number of bits increases the computers word size and address space which increases the processing power in polynomial time. Increasing the number of qubits in a quantum computer increases processing power in exponential time. Unlike a traditional computer, doubling the size of a quantum more than doubles computational power. The increase in computational power is due to two major factors, 1) unique superposition properties of the qubit, 2) higher dimensional algorithms applicable specific problem spaces. Quantum computers provide a different operational computing model when compared to a traditional computer. Rather than serialized approach to computing using logic gates, lasers and radio waves interfere with each other and operate across many qubits simultaneously. In some qubits, interference is constructive and in others interference is destructive. The design of the quantum computer and algorithms seek to reinforce constructive interference patterns that produce the desired results. I realize this answer is not satisfactory for everyone. Take a look at the presentation materials in the links at the of the post. Also take a look at, The Limits of Quantum article.
Quantum computers not likely to replace traditional computer
Quantum computers are fast at solving specific problems where an algorithm exists. Quantum computers are not necessarily fast at solving all problems. It’s unlikely a quantum computer will replace your desktop; however, if a quantum computer could be made small enough it could make an addition to your desktop for specialized functions (e.g., 3D graphics).
Implications for web browser security
A quantum algorithm exists for finding large prime numbers, Shore’s Algorithm. Web browser security is predicated on the fact that large prime numbers are difficult to factor. A quantum computer along with Shore’s Algorithm can factor primes fast. However, the state of the art in quantum computers today is about 9-qubits. According to Professor Dan Boneh, we don’t need to be concerned about quantum computers cracking browser security until quantum computers reach around 100-qubits.
Browser security in a post-quantum computing world
Professor Boneh elaborated, post-quantum computing encryption algorithms remain an area of interest. Algorithms that are useful in a post-quantum world favor smaller primes within higher dimensional number spaces(>1024). A research paper, Post-Quantum Key Exchange – A New Hope provides details.
TLS-RAR for auditing/monitoring SSL/TLS connections
A new protocol has been developed to monitor SSL/TLS. TLS-RAR does not require terminating the SSL/TLS connection and establishing a new connection to the end-point. Instead TLS-RAR works by dividing TLS connections into multiple epochs. As a new epoch is established, between client and server, a new TLS session key is negotiated. Meanwhile, the TLS session key for old epochs is provided to the observer which may be an auditor or monitoring tool. In this way the observer has access to view old TLS epoch information. The observer cannot view or alter information from the current epoch. Data integrity and confidentiality between client and server is maintained. Some of the advantages, no changes to the client are required(no new roots to add), and support for current TLS/SSL libraries. This means TLS-RAR is compatible with a host of IoT technologies and components already deployed.
Session Media from the Forum
The following links provide access to session materials throughout the form.